US20020176624A1  Method of isomorphic singular manifold projection still/video imagery compression  Google Patents
Method of isomorphic singular manifold projection still/video imagery compression Download PDFInfo
 Publication number
 US20020176624A1 US20020176624A1 US09/745,354 US74535400A US2002176624A1 US 20020176624 A1 US20020176624 A1 US 20020176624A1 US 74535400 A US74535400 A US 74535400A US 2002176624 A1 US2002176624 A1 US 2002176624A1
 Authority
 US
 United States
 Prior art keywords
 frame
 image
 method
 compression
 polynomial
 Prior art date
 Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
 Abandoned
Links
 238000007906 compression Methods 0 abstract claims description title 186
 230000001131 transforming Effects 0 claims description 22
 230000000875 corresponding Effects 0 claims description 19
 238000004422 calculation algorithm Methods 0 claims description 13
 230000002068 genetic Effects 0 claims description 4
 230000001537 neural Effects 0 claims description 3
 238000000034 methods Methods 0 description 90
 241000282414 Homo sapiens Species 0 description 23
 239000011159 matrix materials Substances 0 description 13
 230000002829 reduced Effects 0 description 12
 230000001603 reducing Effects 0 description 12
 238000004458 analytical methods Methods 0 description 11
 238000004891 communication Methods 0 description 11
 238000006722 reduction reaction Methods 0 description 11
 230000002123 temporal effects Effects 0 description 11
 241000283220 Odobenus rosmarus Species 0 description 10
 230000000007 visual effect Effects 0 description 10
 230000019771 cognition Effects 0 description 9
 210000004279 Orbit Anatomy 0 description 8
 244000052616 bacterial pathogens Species 0 description 8
 239000000872 buffers Substances 0 description 8
 238000003860 storage Methods 0 description 8
 238000006467 substitution reaction Methods 0 description 8
 239000003570 air Substances 0 description 7
 239000003086 colorant Substances 0 description 6
 239000000727 fractions Substances 0 description 6
 238000000926 separation method Methods 0 description 6
 238000000844 transformation Methods 0 description 6
 230000001965 increased Effects 0 description 5
 238000003780 insertion Methods 0 description 5
 210000001508 Eye Anatomy 0 description 4
 101700084288 NS4 family Proteins 0 description 4
 101700016239 SPIKE family Proteins 0 description 4
 101700043803 VP6 family Proteins 0 description 4
 101700060822 VP6A family Proteins 0 description 4
 101700038373 VP6B family Proteins 0 description 4
 230000001419 dependent Effects 0 description 4
 230000002708 enhancing Effects 0 description 4
 239000000284 extracts Substances 0 description 4
 230000004438 eyesight Effects 0 description 4
 239000002609 media Substances 0 description 4
 230000011218 segmentation Effects 0 description 4
 230000003595 spectral Effects 0 description 4
 241000282412 Homo Species 0 description 3
 210000001525 Retina Anatomy 0 description 3
 210000000857 Visual Cortex Anatomy 0 description 3
 238000009825 accumulation Methods 0 description 3
 230000015572 biosynthetic process Effects 0 description 3
 230000002079 cooperative Effects 0 description 3
 230000002596 correlated Effects 0 description 3
 230000003247 decreasing Effects 0 description 3
 238000000605 extraction Methods 0 description 3
 239000010408 films Substances 0 description 3
 238000005755 formation Methods 0 description 3
 210000002569 neurons Anatomy 0 description 3
 230000004044 response Effects 0 description 3
 238000007514 turning Methods 0 description 3
 241000270299 Boa Species 0 description 2
 210000004556 Brain Anatomy 0 description 2
 210000003986 Cell, Retinal Photoreceptor Anatomy 0 description 2
 210000000887 Face Anatomy 0 description 2
 240000004958 Talinum portulacifolium Species 0 description 2
 210000004027 cells Anatomy 0 description 2
 238000010276 construction Methods 0 description 2
 238000009826 distribution Methods 0 description 2
 230000000694 effects Effects 0 description 2
 238000005516 engineering processes Methods 0 description 2
 238000003384 imaging method Methods 0 description 2
 230000001788 irregular Effects 0 description 2
 239000000203 mixtures Substances 0 description 2
 210000003370 receptor cell Anatomy 0 description 2
 239000007787 solids Substances 0 description 2
 101700038984 CSGA family Proteins 0 description 1
 241000282828 Camelus bactrianus Species 0 description 1
 101700076960 M1 family Proteins 0 description 1
 210000000214 Mouth Anatomy 0 description 1
 210000001331 Nose Anatomy 0 description 1
 101700008985 P7 family Proteins 0 description 1
 101700040380 VNS2 family Proteins 0 description 1
 101700002371 VP8 family Proteins 0 description 1
 230000000996 additive Effects 0 description 1
 239000000654 additives Substances 0 description 1
 229930002945 alltransretinaldehyde Natural products 0 description 1
 230000006399 behavior Effects 0 description 1
 238000005452 bending Methods 0 description 1
 230000000903 blocking Effects 0 description 1
 230000003139 buffering Effects 0 description 1
 239000006227 byproducts Substances 0 description 1
 230000015556 catabolic process Effects 0 description 1
 239000003518 caustic Substances 0 description 1
 238000006243 chemical reaction Methods 0 description 1
 239000002131 composite material Substances 0 description 1
 230000004059 degradation Effects 0 description 1
 238000006731 degradation Methods 0 description 1
 230000004069 differentiation Effects 0 description 1
 239000000686 essences Substances 0 description 1
 238000009472 formulation Methods 0 description 1
 SWQJXJOGLNCZEYBJUDXGSMSAN helium3 Chemical compound data:image/svg+xml;base64,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 data:image/svg+xml;base64,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 [3He] SWQJXJOGLNCZEYBJUDXGSMSAN 0 description 1
 230000001976 improved Effects 0 description 1
 238000002372 labelling Methods 0 description 1
 239000010410 layers Substances 0 description 1
 239000011133 lead Substances 0 description 1
 230000036961 partial Effects 0 description 1
 230000000737 periodic Effects 0 description 1
 239000002243 precursor Substances 0 description 1
 238000004321 preservation Methods 0 description 1
 239000011604 retinal Substances 0 description 1
 230000002207 retinal Effects 0 description 1
 238000005204 segregation Methods 0 description 1
 230000035945 sensitivity Effects 0 description 1
 230000001953 sensory Effects 0 description 1
 230000003068 static Effects 0 description 1
 230000001702 transmitter Effects 0 description 1
Images
Classifications

 G—PHYSICS
 G06—COMPUTING; CALCULATING; COUNTING
 G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
 G06T9/00—Image coding
 G06T9/001—Modelbased coding, e.g. wire frame
Abstract
Methods and apparatuses for still image compression, video compression and automatic target recognition are disclosed. The method of still image compression uses isomorphic singular manifold projection whereby surfaces of objects having singular manifold representations are represented by best match canonical polynomials to arrive at a model representation. The model representation is compared with the original representation to arrive at a difference. If the difference exceeds a predetermined threshold, the difference data are saved and compressed using standard lossy compression. The coefficients from the best match polynomial together with the difference data, if any, are then compressed using lossless compression. The method of motion estimation for enhanced video compression sends I frames on an “asneeded” basis, based on comparing the error between segments of a current frame and a predicted frame. If the error exceeds a predetermined threshold, which can be based on program content, the next frame sent will be an I frame. The method of automatic target recognition (ATR) including tracking, zooming, and image enhancement, uses isomorphic singular manifold projection to separate texture and sculpture portions of an image. Soft ATR is then used on the sculptured portion and hard ATR is used on the texture portion.
Description
 The present invention relates to image compression systems, and in particular relates to an image compression system which provides hypercompression.
 BACKGROUND OF THE INVENTION
 Image compression reduces the amount of data necessary to represent a digital image by eliminating spatial and/or temporal redundancies in the image information. Compression is necessary in order to efficiently store and transmit still and video image information. Without compression, most applications in which image information is stored and/or transmitted would be rendered impractical or impossible.
 Generally speaking, there are two types of compression: lossless and lossy. Lossless compression reduces the amount of image data stored and transmitted without any information loss, i.e., without any loss in the quality of the image. Lossy compression reduces the amount of image data stored and transmitted with at least some information loss, i.e., with at least some loss of quality of the image.
 Lossy compression is performed with a view to meeting a given available storage and/or transmission capacity. In other words, external constraints for a given system may define a limited storage space available for storing the image information, or a limited bandwidth (data rate) available for transmitting the image information. Lossy compression sacrifices image quality in order to fit the image information within the constraints of the given available storage or transmission capacity. It follows that, in any given system, lossy compression would be unnecessary if sufficiently high compression ratios could be achieved, because a sufficiently high compression ratio would enable the image information to fit within the constraints of the given available storage or transmission capacity without information loss.
 The vast majority of compression standards in existence today relate to lossy compression. These techniques typically use cosinetype transforms like DCT and wavelet compression, which are specific types of transforms, and have a tendency to lose high frequency information due to limited bandwidth. The “edges” of images typically contain very high frequency components because they have drastic gray level changes, i.e., their dynamic range is very large. Edges also have high resolution. Loss of edge information is undesirable because resolution is lost as well as high frequency information. Furthermore, human cognition of an image is primarilydependent upon edges or contours. If this information is eliminated in the compression process, human ability to recognize the image decreases.
 Fractal compression, though better than most, suffers from high transmission bandwidth requirements and slow coding algorithms. Another type of motion (video) image compression technique is the ITUrecommended H.261 standard for videophone/videoconferencing applications. It operates at integer multiples of 64 kbps and its segmentation and model based methodology splits an image into several regions of specific shapes, and then the contour and texture parameters representing the region boundaries and approximating the region pixels, respectively, are encoded A basic difficulty with the segmentation and modelbased approach is low image quality connected with the estimation of parameters in 3D space in order to impart naturalness to the 3D image. The shortcomings of this technique are obvious to those who have used videophone/videoconferencing applications with respect particularly to MPEG video compression.
 Standard MPEG video compression is accomplished by sending an “I frame” representing motion every fifteen frames regardless of video content. The introduction of I frames asynchronously into the video bitstream in the encoder is wasteful and introduces artifacts because there is no correlation between the I frames and the B and P frames of the video. This procedure results in wasted bandwidth. Particularly, if an I frame has been inserted into B and P frames containing no motion, bandwidth is wasted because the I frame was essentially unnecessary yet, unfortunately, uses up significant bandwidth because of its full content. On the other hand, if no I frame is inserted where there is a lot of motion in the video bitstream, such overwhelming and significant errors and artifacts are created that bandwidth is exceeded. Since the bandwidth is exceeded by the creation of these errors, they will drop off and thereby create the much unwanted blocking effect in the video image. In the desired case, if an I frame is inserted where there is motion (which is where an I frame is desired and necessary) the B and P frames will already be correlated to the new motion sequence and the video image will be satisfactory. This, however, happens only a portion of the time in standard compression techniques like MPEG. Accordingly, it would be extremely beneficial to insert I frames only where warranted by video content.
 The compression rates required in many applications including tactical communications are extremely high as shown in the following example making maximal compression of critical importance. Assuming 512^{2 }number of pixels, 8bit gray level, and 30 Hz fullmotion video rate, a bandwidth of 60 Mbps is required. To compress data into the required data rate of 128 kbps from such a full video uncompressed bandwidth of 60 Mbps, a 468:1 still image compression rate is required. The situation is even more extreme for VGA fullmotion video which requires 221 Mbps and thus a 1726:1 motion video compression rate. Such compression rates, of course, greatly exceed any compression rate achievable by state of the art technology for reasonable PSNR (peak signal to noise ratio) values of approximately 30 dB. For example, the fourth public release of JPEG has only a 30:1 compression rate and the image has many artifacts due to a PSNR of less than 20 dB, while H320 has a 300:1 compression ratio for motion and still contains many still/motion image artifacts.
 The situation is even more stringent for continuity of communication when degradation of power budget or multipath errors of wireless media further reduce the allowable data rate to far below 128 kbps. Consequently, state of the art technology is far from providing multimedia parallel channelization and continuity data rates at equal to or lower than 128 kbps.
 Very high compression rates, high image quality, and low transmission bandwidth are critical to modem communications, including satellite communications, which require fullmotion, high resolution, and the ability to preserve highquality fidelity of digital image transmission within a small bandwidth communication channel (e.g. T1). Unfortunately, due to the above limitations, state of the art compression techniques are not able to transmit high quality video in realtime on a bandlimited communication channel. As a result, it is evident that a compression technique for both still and moving pictures that has a very high compression rate, high image quality, and low transmission bandwidth and a very fast decompression algorithm would be of great benefit. Particularly, a compression technique having the above characteristics and which preserves high frequency components as well as edge resolution would be particularly useful.
 In addition to transmission or storage of compressed still or moving images, another area where the state of the art is unsatisfactory is in automatic target recognition (ATR). There are numerous applications, both civilian and military, which require the fast recognition of objects or humans amid significant background noise. Two types of ATR are used for this purpose, soft ATR and hard ATR. Soft ATR is used to recognize general categories of objects such as tanks or planes or humans whereas hard ATR is used to recognize specific types or models of objects within a particular category. Existing methods of both soft and hard ATR are Fourier transformbased. These methods are lacking in that Fourier analysis eliminates desired “soft edge” or contour information which is critical to human cognition. Improved methods are therefore needed to achieve more accurate recognition of general categories of objects by preserving critical “soft edge” information yet reducing the amount of data used to represent such objects and thereby greatly decrease processing time, increase compression rates, and preserve image quality.
 The present invention is based on Isomorphic Singular Manifold Projection (ISMP) or Catastrophe Manifold Projection (CMP). This method is based on Newtonian polynomial space and characterizes the images to be compressed with singular manifold representations called catastrophes. The singular manifold representations can be represented by polynomials which can be transformed into a few discrete numbers called “datery” (number data that represent the image) that significantly reduce information content. This leads to extremely high compression rates (CR) for both still and moving images while preserving critical information about the objects in the image.
 In this method, isomorphic mapping is utilized to map between the physical boundary of a 3D surface and its 2D plane. A projection can be represented as a normal photometric projection by adding the physical parameters, B (luminance) to generic geometric parameters (X,Y). This projection has a unique 3D interpretation in the form of a “canonical singular manifold”. This manifold can be described by a simple polynomial and therefore compressed into a few discrete numbers resulting in hyper compression. In essence, any image is a highly correlated sequence of data. The present invention “kills” this correlation, and image information in the form of a digital continuum of pixels almost disappears. All differences in 2D “texture” connected with the 2D projection of a 3D object are “absorbed” by a contour topology, thus preserving and emphasizing the “sculpture” of the objects in the image. This allows expansion with good fidelity of a 2D projection of a real 3D object into an abstract (mathematical) 3D object and is advantageous for both still and video compression and automatic target recognition.
 More particularly, using catastrophe theory, surfaces of objects may be represented in the form of simple polynomials that have singlevalued (isomorphic) inverse reconstructions. According to the invention, these polynomials are chosen to represent the surfaces and are then reduced to compact tabulated normal form polynomials which comprise simple numbers, i.e., the datery, which can be represented with very few bits. This enables exceptionally high compression rates because the “sculpture” characteristics of the object are isomorphically represented in the form of simple polynomials having singlevalued inverse reconstructions. Preservation of the “sculpture” and the soft edges or contours of the object is critical to human cognition of the image for both still and video image viewing and ATR. Thus, the compression technique of the present invention provides exact representation of 3D projection edges and exact representation of all the peculiarities of moving (rotating, etc.) 3D objects, based on a simple transition between still picture representation to moving pictures.
 In a preferred embodiment the following steps may be followed to compress a still image using isomorphic singular manifold projections and highly compressed datery. The first step is to subdivide the original image, I_{O}, into blocks of pixels, for example 16×16 or other sizes. These subdivisions of the image may be fixed in size or variable. The second step is to create a “canonical image” of each block by finding a match between one of fourteen canonical polynomials and the intensity distribution for each block or segment of pixels. The correct polynomial is chosen for each block by using standard merit functions. The third step is to create a model image, I_{M}, or “sculpture” of the entire image by finding connections between neighboring local blocks or segments of the second step to smooth out intensity (and physical structure to some degree). The fourth step is to recapture and work on the delocalized high frequency content of the image, i.e., the “texture”. This is done by a subtraction of the model image, I_{M}, generated during the third step from the original segmented image, I_{O}, created during the first step. A preferred embodiment of this entire still image compression process will be discussed in detail below.
 Optimal compression of video and other media containing motion may be achieved in accordance with the present invention by inserting I frames based on video content as opposed to at fixed intervals (typically every 15 frames) as in the in prior art motion estimation methods. In accordance with the motion estimation techniques of the present invention, the errors between standard “microblocks” or segments of the current frame and a predicted frame are not only sent to the decoder to reconstruct the current frame, but, in addition, are accumulated and used to determine the optimal insertion points for I frames based on video content. Where the accumulated error of all the microblocks for the current frame exceeds a predetermined threshold which itself is chosen based upon the type of video (action, documentary, nature, etc.), this indicates that the next subsequent frame after the particular frame having high accumulated error should be an I frame. Consequently, in accordance with the present invention, where the accumulated errors between the microblocks or segments of the current frame and the predicted frame exceed the threshold, the next subsequent frame is sent as an I frame which starts a new motion estimation sequence. Consequently, I frame insertion is content dependent which greatly improves the quality of the compressed video.
 The I frames inserted in the above compression technique may first be compressed using standard DCT based compression algorithms or the isomorphic singular manifold projection (ISMP) still image compression technique of the present invention for maximal compression. In either case, the compression techniques used are preferably MPEG compatible.
 Additionally, using the motion estimation technique compression of the present invention, compression ratios can be dynamically updated from frame to frame utilizing the accumulated error information. The compression ratio may be changed based on feedback from the receiver and, for instance, where the accumulated errors in motion estimation are high, the compression ratio may be decreased, thereby increasing bandwidth of the signal to be stored. If, on the other hand, the error is low, the compression ratio can be increased, thereby decreasing bandwidth of the signal to be stored.
 Because the present invention is a 3D nonlinear technique that produces high level descriptive image representation using polynomial terms that can be represented by a few discrete numbers or datery, it provides much higher image compression than MPEG (greater than 1000:1 versus 100:1 in MPEG), higher frame rate (up to 60 frames/sec versus 30 frames/sec in MPEG), and higher picture quality or peak signal to noise ratio (PSNR greater than 32 dB versus PSNR greater than 23 dB in MPEG). Consequently, the compression technique of the present invention can provide more video channels than MPEG for any given channel bandwidth, video frame rate, and picture quality.
 FIG. 1A illustrates a singularity called a “fold”; and FIG. 1B illustrates a singularity called a “cusp”;
 FIG. 2A illustrates “Newton Diagram Space” and contains “monoms” and polynomials;
 FIG. 2B illustrates the application of Newton diagram space in the context of ISMP Theory. Canonical and normal forms, etc.;
 FIG. 3A depicts a fold and FIG. 3B depicts tuck;
 FIG. 4A illustrates a reflection from a manifold;
 FIG. 4B illustrates the reflection from a manifold depending upon angle θ;
 FIG. 4C illustrates the projection on a display;
 FIG. 5 illustrates a cylinder with constant luminance dependence;
 FIG. 6 illustrates an evolvement for F=X^{3};
 FIGS. 7A and 7B illustrate a reflector of a group for three mirrors R^{2};
 FIG. 8 illustrates a smooth curve projection, representing movement and a physical object, and a catastrophic frame change, and positionally zoom camera changes;
 FIG. 9 is a abbreviated flow chart of the inventive ISMP still image compression method;
 FIG. 10 is a detailed flow chart of the inventive ISMP compression method in accordance with the present invention;
 FIG. 11A illustrates an original image with an enlarged edge contour;
 FIG. 11B shows a 2D CCD image of the enlarged edge contour;
 FIG. 11C illustrates a model surface of the original edge contour in accordance with the present invention;
 FIG. 12 is an illustration of a segment of an original frame in accordance with the present invention;
 FIG. 13 is an illustration of connecting segments of a frame in accordance with the present invention;
 FIG. 14 illustrates subtracting I_{M }from I_{O }in accordance with the present invention;
 FIG. 15 is a flow chart of the decoding process for ISMP compression in accordance with the present invention;
 FIG. 16 is a flow chart of the motion estimation process in accordance with the present invention;
 FIG. 17 is a circuit schematic of the hardware uses for motion estimation in accordance with the present invention;
 FIG. 18 is a flow chart of the error accumulation method of the present invention; and
 FIG. 19 is a table showing the results of data communication with varying data rates in accordance with the present invention.
 The following is a brief introduction to Catastrophe Theory which may be helpful in understanding novel compression methodology of the present invention. Further discussion may be found in B. I. Arnold,Catastrophe Theory, SpringerVerlag 1992, which is incorporated by reference herein.
 Catastrophes are abrupt changes arising as a sudden response of a system to a smooth change in external conditions. In order to understand catastrophe theory, it is necessary to understand Whitney's singularity theory. A mapping of a surface onto a plane associates to each point of the surface a point of the plane. If a point on the surface is given coordinates (x_{1}, x_{2}) on the surface, and a point on the plane is given coordinates (y_{1}, y_{2}), then the mapping is given by a pair of functions y_{1}=f_{1}(x_{1}, x_{2}) and y_{2}=f_{2}(x_{1}, x_{2}). The mapping is said to be smooth if these functions are smooth (i.e., are differentiable a sufficient number of times, such as polynomials for example). Mappings of smooth surfaces onto a plane exist everywhere. Indeed, the majority of objects surrounding us are bounded by smooth surfaces. The visible contours of bodies are the projections of their bounding surfaces onto the retina of the eye. By examining the objects surrounding us, for instance, people's faces, the singularities of visible contours can be studied. Whitney observed that generically (for all cases bar some exceptional ones) only two kinds of singularities are encountered. All other singularities disintegrate under small movements of the body or of the direction of projection, while these two types are stable and persist after small deformations of the mapping.
 An example of the first kind of singularity, which Whitney called a fold, is the singularity arising at equatorial points when a sphere is projected onto a plane such as shown in FIG. 1A. In suitable coordinates, this mapping is given by the formulas
 y _{1} =x _{1} ^{2} , y _{2} =x _{2}
 The projections of surfaces of smooth bodies onto the retina have just such a singularity at generic points, and there is nothing surprising about this. What is surprising is that besides the singularity, the fold, we encounter everywhere just one other singularity, but it is practically never noticed.
 The second singularity was named the cusp by Whitney, and it arises when a surface like that in FIG. 1B is projected onto a plane. This surface is given by the equation
 y _{1} =x _{1} ^{3} +x _{1} x _{2} , y _{2} =x _{2}
 with respect to spatial coordinates (x_{1}, x_{2}, y_{1}) and projects onto the horizontal plane (x_{2}, y_{1}).
 On the horizontal projection plane, one sees a semicubic parabola with a cusp (spike) at the origin. This curve divides the horizontal plane into two parts: a smaller and a larger one. The points of the smaller part have three inverse images (three points of the surface project onto them), points of the larger part only one, and points on the curve, two. On approaching the curve from the smaller part, two of the inverse images (out of three) merge together and disappear (here the singularity is a fold), and on approaching the cusp all three inverse images coalesce.
 Whitney proved that the cusp is stable, i.e., every nearby mapping has a similar singularity at an appropriate nearby point (that is, a singularity such that the deformed mapping, in suitable coordinates in a neighborhood of the point mentioned, is described by the same formulas as those describing the original mapping in a neighborhood of the original points). Whitney also proved that every singularity of a smooth mapping of a surface onto a plane, after an appropriate small perturbation, splits into folds and cusps. Thus, the visible contours of generic smooth bodies have cusps at points where the projections have cusp singularities, and they have no other singularities. These cusps can be found in the lines of every face or object. Since smooth mappings are found everywhere, their singularities must be everywhere also, and since Whitney's theory gives significant information on singularities of generic mappings, this information can be used to study large numbers of diverse phenomena and processes in all areas of science. This simple idea is the whole essence of catastrophe theory.
 The following glossary is a useful aid to understanding catastrophe theory because many of the terms used to describe it are uncommon in mathematics.
 2D Cartesian (Plane) Coordinates refer to standard (u, v) coordinates that describe a plane projection.
 2D Generalized Coordinates: (ξ,ν) describe a system through a minimum number of geometrical coordinates (i.e., a number of degrees of freedom). These are usually curvilinear local coordinates, which belong to a specific surface in the vicinity of some point (i.e., origin of coordinates).
 3D Cartesian Coordinates refer to (x,y,z) describing a common surface in 3D: F (x,y,z)=0.
 3D Cartesian (Hyperplane) Coordinates, are: (u, v, w), where (u, v) are 2D (plane) Cartesian coordinates; w is a third, new physical coordinate, related to luminance (B) and describing a “gray level” scale color scale.
 Arnold (Vladimir) is a Russian mathematician, who is a major contributor to catastrophe theory.
 Arnold Theorem (Local Isomorphism) A family of transformations can transform any given mapping into a set of canonical transformations by using smooth substitutions of coordinates. The Arnold theorem defines local isomorphism in a sense that defines a class of locally isomorphic functions.
 1. Arnold proved that Thom's theory can be represented in terms of group theory. 2. He also introduced an elegant theory for construction of the canonical form of singularities as they apply to wave front propagation in Lagrangian mechanics. 3. Furthermore, Arnold introduced methods based on using algebra of vector fields
$v=\sum _{i}\ue89e{R}_{i}\ue89e\frac{\partial \text{\hspace{1em}}}{\partial x},$  where R_{i }is a polynomial.
 and introduced a method of spectral series for reduction of arbitrary functions to normal form. 4. Finally, he introduced classification of singularities and a method that described how to determine any type of singularity within a list of singularities.
 Canonical Form is a generic mathematical term that can be defined in various ways. In the specific context of the Arnold theorem, the canonical form is the simplest polynomial, with the highest degree of monoms within th e normal form area, representing a given type of catastrophe. The canonical form is represented by a segmented line in a Newton diagram.
 Canonical Transformation permits transformation of real surface form (such as F (u, v, w)=0) into canonical form (i.e., superposition of Morse form and singular residuum, or Thom form) into two blocks of socalled canonical coordinates: regular and catastrophic.
 Catastrophe (a term invented by Montel) is a specific manifold mapping feature by which some points lying in the projection plane can abruptly change location in manifold. More philosophically, it “describes the emergence of discrete structures from the typical surface described in the platform of continuum.”
 Catastrophes, Critical Number in 3D (for mapping a generic surface onto a plane) is only two (2): fold and cusp (tuck). Using these two catastrophes is sufficient for static still imagery.
 Catastrophes, Total Number in 3D (for mapping a generic surface onto a plane) is fourteen (14). Only “fold”catastrophe does not have degenerate points; all the (13) others have. Using all 14 catastrophes is necessary in hypercompression if we consider dynamic imagery (or video).
 Catastrophic Manifold Projection is a fundamental concept of “3D into 3D” mapping, leading to hypercompression. This is diffeomorphic mapping, including geometrical coordinates (2D generalized, and 2D plane), as well as a fourth “photometric coordinate”.
 Catastrophic Manifold Projection (CMP) Law is mapping:
 (ξ, ν, B)←→(u, v, w)
 Thus, the CMP is “3D into 3D” mapping, with two types of coordinates: “geometrical” (ξ, ν); (u, v), and “photometric” (B, w).
 The Critical Point is a point at which the rank of Jacobian is less than maximal (examples are maxima, minima, and bending points)
 Datery result from a novel mathematical procedure leading to a tremendous compression ratio; instead of describing some surface by a continuum, we describe this singular manifold by a few low even numbers, i.e., datery. Therefore, during hypercompression, the surface as continuum “disappears”, leaving typical data (such as computer data).
 In a Degenerate Critical Point, the rank of Jacobian is a less than maximal rank minus one. This point can be a critical point of cusp catastrophes, for example.
 Discrete Structures are singular manifolds that can be described by a set of discrete, usually even, data (e.g., (2,5,1,3)), leading to datery instead of description by a continuum of points (such as F(x,y,z)=0). Such discrete structures (which are, in fact, continuums, but are still described by discrete sets), are typically referred to as singularities, bifurcation, or catastrophes.
 Diffeomorphism is a stronger term than isomorphism (or homeomorphism), and means not only isomorphism, but also smooth mapping.
 Discrete Structures are singular manifolds that can be described by a set of discrete data (e.g., (2,5,1,3)), leading to datery, instead of description by continuum of points (such as F(x,y,z)=0). These discrete structures (which are, in fact, continuums, but are still described by discrete sets), are typically referred to as singularities, bifurcation, or catastrophes.
 Field, a subset of a ring. (All nonzero field elements generate a group, by multiplication.) For example, the differential operator can be an element of a field.
 Generalized Coordinates, or, more precisely, generalized coordinates of Lagrange, are such “natural” coordinates in solid state mechanics that their number is precisely equal to the number of a body's degrees of freedom: (ξ, ν, η, . . . )
 A Generic Surface, in the context of the CMP method, is a mathematical surface which, within infinitesimal changes, does not have the same tangent (or projection line) for more than two points along any curve lying on this surface. In other words, a surface is generic if small changes to the surface do not lead to changes in singularities (such as splitting into less complex ones.) Physical surfaces are almost always generic because of noise tolerance.
 A Group is the simplest set in mathematical models, with only a single operation.
 Hypercompression is a specific compression term which provides a datery (i.e., “stripping” a continuum surface into its discrete representation). This is possible for the surface locality in the form of catastrophe.
 Ideal is another subset of a ring. A subset of the ring is an ideal if this subset is a subgroup of the ring by summation. In the context of the Arnold theorem, this summation group is a set of all monoms that lie above the canonical form segmented line.
 Isomorphic Singular Manifold Projection—see CMP.


 describes a partial derivative.
 Landau (Lev) was a Russian theoretical physicist, who won the Nobel Prize for superfluidity of the isotope helium He_{3}. He systematically applied the catastrophe theory approach before this theory was mathematically formulated^{[3]}.
 The Landau Method, applied in the secondorder phase transition, applies Thom's lemma in the form of the Taylor series, including only “important” physical terms.
 Lie Algebra is an algebra belonging to the Lie groups with a binary operation (commutator).
 A Lie Group is a group whose generator is an infinitesimal operator.
 Locally Isomorphic functions have the same singular residuum (see Thom's lemma); thus, they can be compressed identically for “soft edges”, or “object boundaries”.
 Manifold is a mathematical surface (curve or point) defined locally by a system of equations through local “canonical” coordinates, also called curvilinear (natural) coordinates, or generalized coordinates of Lagrange (known as generalized coordinates for short).
 Mapping is a transformation in which
 u=f(ξ, ν) (3)
 v=g(ξ, ν)
 and vice versa. Mapping is smooth if functions f and g are smooth (i.e., differentiable a “sufficient” number of times: the highest level of “sufficient” differentiation is equivalent to the highest power of a polynomial describing a given manifold.
 A Monom is a point in Newton diagram space, describing a given polynomial term. For example, term: x^{3}y is equivalent to the monom (3,1), a point in a Newton diagram (FIGS. 2A and 2B).
 Morse was a French mathematician whose work was a precursor of catastrophe theory. At the beginning of the nineteenth century, he generalized a number of differential geometry theorems into a general class of generic surfaces.
 Morse lemma: In the vicinity of a nondegenerate critical point of specific manifold mapping, a function describing specific manifold mapping in generalized coordinates can be reduced to a quadratic form.
 A Newton Diagram is discrete (Cartesian) 2D “point” space defmed in such a way that the xaxis and the yaxis describe xpolynomial and ypolynomial power, respectively. For example, the x^{2}ypolynomial element is equivalent to point (2,1) in Newton table space. In this Newton diagram space, a given polynomial that is always normalized (i.e., with unit coefficients of x^{2}+xy, and not x^{2}+3xy) is described by a segmented line. See FIGS. 3A and 3B.
 Nondegenerate Critical Point For this point, only one row of Jacobian is equal to zero (this point can be maximum or minimum, as referred to in the Morse lemma).
 Normal Form is a set of monoms bounded by a canonical form segmented line (including the monoms of canonical form).
 A Ring is the second most complex set in mathematical models, with two operations. Ring subsets can be field and ideal.
 Stable Catastrophes are always two: fold and cusp (tuck). These cannot be “easily” transferred to another catastrophe by infinitesimal transformation although others can be. See FIGS. 1 and 2.
 A Spectral Series is a method of sequential approximation (proposed by Arnold) that allows reduction of all catastropheequivalent polynomials to the canonical form, representing a given type of catastrophe.
 Thom (Rene) was a French mathematician, considered to be the “father” of catastrophe theory (1959).
 Thom's lemma is a fundamental theorem of catastrophe theory in general and the ISMP in particular, as a generalization of the Morse lemma for degenerate critical points. It claims that, in such a case, the algebraic form describing a surface can no longer be only quadratic, but consists of a quadratic form (as in the Morse lemma) and an additive singular residuum:
$f=\sum _{i=n}^{s}\ue89e\text{\hspace{1em}}\ue89e{k}_{i}\ue89e{\alpha}_{1}^{2}+g\ue89e\left({\alpha}_{s+n},\dots \ue89e\text{\hspace{1em}}\ue89e{\alpha}_{n}\right),{k}_{i}=\pm 1$  Thesenormalized coordinates are also separated into two parts: nongenerate point coordinates (NPC) (i=1,2, . . . ,s) and degenerate point coordinates (DPC) (i=s+1,s+2, . . . ,n). In the residuum function g(α_{s+1}, . . . ,α_{n}), the first and secondorder differentials vanish: dg=d^{2}g=0. Functions with the same g belong to a set of stable equivalent functions, or are locally isomorphic (Arnold).
 The Thom Statement declares that there is a finite number of catastrophes (14) in 3D space.
 A Vector Field is a representation whose element provides a shift of polynomials in the Newton diagram (this shift does not need to be a translation).
 Whitney (M.) (1955) was an American mathematician, and a major contributor to catastrophe theory. His major achievements were in studying mapping from surface to plane.
 Whitney Theorem (Two Stable Catastrophes): The local normal form of the singularities of typical stable mappings from 2D manifolds (in 3D) to a plane can be either fold or cusp only. (Stable Mapping): Every singularity of smooth mapping of a surface onto a plane after an appropriate small perturbation splits into stable catastrophes only (fold and cusp). This theorem is applied in CMP hypercompression into still imagery.
 The following references are referred to in the text that follows and are hereby incorporated by reference.
 1. M. Born, E. Wolf,Principles of Optics, Pergamon Press, 1980.
 2. T. Jannson, “Radiance Transfer Function,” J. Opt. Soc. Am. Vol. 70, No. 12, 1980] pp. 15441549.
 3. V. I. Arnold,Catastrophe Theory SpringerVerlag, NY, 1992.
 4. V. I. Arnold,Singularities of Caustics and Wave Fronts, Mathematics and Its Applications (Sovien Series) Vol. 62, Kluwer Academic Publisher, 1990.
 5. V. I. Arnold, The Theory of Singularity and its Applications, Academia Nazionale dei Lincei, 1993.
 6. V. I. Arnold, S. M. GuseinZade, A. N. Varchenko,Singularities of Differential Mapping, Birkhäuser, BostonBaselBerlin, 1988.
 7. R. Gilmore,Catastrophe Theory for Scientists and Engineers, John Wiley & Sons, New York, 1981.
 8. P. Grey,Psychology, Worth Publishers, New York, N.Y., 1991.
 The following are expanded definitions, theorems, and lemmas referred to in the discussion below:
 Critical Point: For a function depending on n variables (ξεR^{n }or ndimensional magnified), a critical point is called nondegenerate if its second differential is a nondegenerate quadratic form. In other words, for this point, only one row of the Jacobian is equal to zero.
 Noncritical Point: In the neighborhood of regular (or noncritical) point transformations of n local coordinates ξ_{i }of a surface into coordinates u_{i }on a mapping plane, the transformation can be written as:
 u _{i} =u _{i}(ξ_{1}, . . . ,ξ_{n}), i=1, . . . ,n.
 In this case, the Jacobian is always nondegenerate. This means that in the vicinity of this point, it is possible to do an isomorphical transformation according to the implicit function theorem:
 ξ_{i}=ξ_{i}(u _{1} , . . . , u _{n}), i=1, . . . ,n
 Morse lemma: In the neighborhood of a nondegenerate critical point, a function may be reduced to its quadratic part, i.e., it may be written into the normal form
 u=ξ _{1} ^{2}− . . . −ξ_{k} ^{2}+ξ_{k} ^{2} _{+1}+ξ (1)
 for a certain local coordinate system (ξ_{1}, . . . , ξ_{n}).
 The meaning of this lemma is as follows: Since the Jacobian of any smooth function f is nonzero in the vicinity of any nondegenerate critical point, differential replacements, such as:
 u _{i} =u _{i}(ξ_{1}, . . . , ξ_{n}) (2)

 At a degenerate critical point, some eigenvalues of the Jacobian matrix are zero. The subspace spanned by the corresponding eigenvalues (ξ_{s+1}, . . . , ξ_{n}) is a critical subspace and has a dimension equal to the corank of the critical point.
 Function f can be written in the form defining Thom's lemma, which is fundamental to the ISMP:
$\begin{array}{cc}{u}_{j}=\sum _{i=1}^{s}\ue89e\text{\hspace{1em}}\ue89e{k}_{i}\ue89e{\xi}_{i}^{2}+g\ue8a0\left({\xi}_{s+1},\dots \ue89e\text{\hspace{1em}},{\xi}_{n}\right)\ue89e\text{\hspace{1em}}\ue89e{k}_{i}=\pm 1& \left(4\right)\end{array}$  where g(ξ_{s+1}, . . . , ξ_{n}) is a function (residuum) for which dg=d^{2}g=0.
 All functions with the same g are called a differential equivalent. The term local isomorphical is used as another description of that class of function.
 Thom's lemma provides a basis for an application mapping algorithm for any surfaces that can be mapped on an image plane.
 Thom's form can be used as a nondegenerate function for image approximation, but using singularities analysis allows extraction of the most important information from an image.
 1st Whitney Theorem. The local normal form of the singularities of typical mappings from twodimensional manifolds to a plane (or to another twodimensional manifold):
$\begin{array}{cc}\mathrm{cusp}\ue89e\{\begin{array}{c}u={\xi}^{3}+\mathrm{\xi \eta}\\ v=\eta \end{array}& \text{(5A)}\\ \mathrm{fold}\ue89e\{\begin{array}{c}u={\xi}^{2}\\ v=\eta \end{array}& \text{(5B)}\\ \mathrm{regular}\ue89e\{\begin{array}{c}u={\xi}^{\text{\hspace{1em}}}\\ v=\eta \end{array}& \text{(5C)}\end{array}$  Stable catastrophes (fold and cusp) are sufficient for still image compression. FIG. 3A depicts a fold and FIG. 3B depicts a cusp.
 2nd Whitney Theorem: Every singularity of a smooth mapping of a surface onto a plane after an appropriate small perturbation splits into folds and cusps.
 Arnold Theorem: There is a family of transformations that can transform any given mapping into a set of canonical transformations by using smooth substitution.

 then, by using smooth diffeomorphic transformation into new “plane” coordinates (u′, v′), we obtain:

$\begin{array}{cc}\begin{array}{c}{u}^{\prime}={a}_{1}\ue89eu+{a}_{2}\ue89ev+{a}_{3}\ue89e{u}^{2}+\dots \\ {v}^{\prime}={b}_{1}\ue89ev+{b}_{2}\ue89eu+{b}_{3}\ue89e{v}^{2}+\dots \\ {u}^{\u2033}={c}_{1}\ue89e{u}^{\prime}+{c}_{2}\ue89e{v}^{\prime}+{c}_{3}\ue89e{u}^{\mathrm{\prime 2}}+\dots \end{array}& \text{(7A)}\\ {v}^{\u2033}={d}_{1}\ue89e{v}^{\prime}+{d}_{2}\ue89e{u}^{\prime}+{d}_{3}\ue89e{v}^{\mathrm{\prime 2}}+\dots & \text{(7B)}\end{array}$  We can obtain
 u=M _{1}(ξ″, η″)+F _{C1}(ξ″, η″)+F _{C2}(ξ″, η″)+ . . . (8)
 v=M _{2}(ξ″, η″)+F _{C3}(ξ″, η″)+ . . .
 where M_{1 }and M_{2 }are Morse forms and F_{C1}, F_{C2}, F_{C3}, . . . are canonical (singular) forms.
 In the Arnold theorem, Thom's lemma is applied in such a sense that we represent a Thom form (as in Eq. 8) by superposition of Morse (smooth) forms (M), and Thom residual forms (F_{C}). The proof of this theorem is based on spectral series reduction to normal forms. This is a local isomorphism (or, more precisely, local diffeomorphism), because each catastrophe is represented by a given canonical form, which, in turn, generates a normal form. Moreover, each catastrophe is represented by only one canonical form. Therefore, while general mapping is usually not isomorphic, in this specific case, Arnold mapping is. The consequence of the Arnold theorem, proven by his students Platonova and Shcherback, is a statement made earlier by Thom: The number of nonequivalent singularities in the projections of generic surfaces in 3D space, defined by the families of rays issuing from different points of space outside the surface, is finite and equal to 14.
 Usually, 3D objects, presented in the form of 2D images, are projections of the following types of objects:
 1. Smooth artificial and natural objects: This category can be described as a projection of idealized surfaces on the image plane. In accordance with the human visual system, we first try to extract objects that can be presented by smooth surfaces (“soft edges”).
 Edges of smooth surfaces: These soft edges, which appear during mapping, are the same as the visible contours of smooth objects. (These objects ideally fit into the proposed approach.)
 2. Sharp joints of objects: One example is the comers of buildings. (These jumps can also be naturally described by the proposed approach.)
 3. Textures of an object: These objects will be described by the proposed method with natural scale parameters (including fractal type textures).
 Formation of an image can be described as light reflection from a general surface. It may be an actual radiation surface (light source, transparent surface, or semitransparent surface) or it may be an opaque reflected surface. We have introduced a photometric projection, so each ray is reflected backward only, in accordance with the radiance projection theorem^{[2]}. The reflection is the highest in the specular direction, and it is monotonically reduced, with an increase of the reflection direction separation from the specular direction, as shown in FIG. 4. This photometric projection approach can be derived from Thom's lemma and the Arnold theorem.
 If a reflection is identified with reflection surface luminance, B, the dependence of B on the direction will depend on the nature of the surface (whether it is smooth or rough). There is no general theory for arbitrary surfaces, although there are two limited cases: Lambert's cosine law, in which B is a constant (isotropic case), and the specular (mirror) reflection, in which incident light is reflected, without distortion, only in the specular direction. In a general case, we have intermediate distribution as seen in FIG. 4A which is a reflection from a manifold,4B which is an explanation of reflection value depending upon angle θ, and 4C which is what would be displayed, for example, on a display. The presented photometric projection has a natural interpretation: the reflection value decreases when the θvalue increases, and vice versa.
 Inverse projection from a 2D plane into the surface of a real object is analyzed. The geometry of this problem (i.e., photometric projection) has been shown in FIG. 4. Now, however, this inverse problem must be formulated in precise mathematical terms, allowing design of a suitable algorithm for hypercompression. To do this, the forward problem of image formation is formulated first.
 In general, image formation can be presented as differential mapping from 4D space (x, y, z, B; where x, y, z are real 3D space coordinates of a point and B is the luminance of that point) to the image plane (u, v, w), where u and v are coordinates of a pixel and w is a color (or gray scale level) of the pixel. The result of mapping the manifold with internal curvilinear (ξ, η) coordinates will be a 3D surface (u, v, w); where u, v are coordinates of the point into the plane and w is the luminance of the point.
 The mapping will be:
 u=f _{1}(x, y, z) (9A)
 v=f _{2}(x, y, z) (9B)
 w=f _{3}(x, y, z, B). (9C)
 or
 u=F _{1}(ξ, η) (10A)
 v=F _{2}(ξ, η) (10B)
 w=F _{3}(ξ, η, B) (10C)
 where f_{1 }and f_{2 }are regular projections of a surface to a plane, f_{3 }is luminance projection, and F_{1}, F_{2}, F_{3 }are their equivalents in the curvilinear coordinate system (ξ, η).
 To formulate the isomorphic singular manifold projection (ISMP) problems by applying Thom's lemma formalism (i.e., canonical form, catastrophes, etc.), one must realize first that the w (B)—dependence is a smooth monotonic one, since both w and B are various forms of luminance, in such a sense that B is the physical luminance, while w is its representation in the form of color (gray level) in the CCD plane. But, smooth dependence does not contain critical points (even nondegenerate ones). Therefore, Thom's (splitting) lemma can be applied to Function (10), in the form:
 w=M(B, ξ, η), or (11A)
 w=M(B, ξ)+g(η), or (11B)
 w=M(B, η)+g(ξ), or (11C)
 w=M(B)+g(ξ, η), (11D)
 where the first function M represents a monotonical function of B or (function without critical points), and g (ξ, η) represents all singularities of projection influencing a gray scale level (color) of a given point (i.e., gfunction represents a singular Thom residuum).
 In order to show this, Function (10C) is expanded into infinite Taylor series, in the vicinity of ξ_{0}, η_{0}, and B_{0}, in the form:
$\begin{array}{cc}w={w}_{0}+\frac{1}{1!}\ue89e\frac{\partial {F}_{3}}{\partial \xi}\ue89e{{\uf603}_{{\xi}_{0}}\ue89e\left(\xi {\xi}_{0}\right)+\frac{1}{1!}\ue89e\frac{\partial {F}_{3}}{\partial \eta}\uf604}_{{\eta}_{0}}\ue89e\left(\eta {\eta}_{0}\right)+\frac{1}{1!}\ue89e\frac{\partial {F}_{3}}{\partial B}\ue89e{{\uf603}_{{B}_{0}}\ue89e\left(B{B}_{0}\right)+\frac{1}{2!}\ue89e\frac{{\partial}^{2}\ue89e{F}_{3}}{\partial {\xi}^{2}}\uf604}_{{\xi}_{0}}\ue89e{\left(\xi {\xi}_{0}\right)}^{2}+\frac{1}{2!}\ue89e\frac{{\partial}^{2}\ue89e{F}_{3}}{\partial {\eta}^{2}}\ue89e{{\uf603}_{{\eta}_{0}}\ue89e{\left(\eta {\eta}_{0}\right)}^{2}+\frac{1}{2!}\ue89e\frac{{\partial}^{2}\ue89e{F}_{3}}{\partial {B}^{2}}\uf604}_{{B}_{0}}\ue89e{\left(B{B}_{0}\right)}^{2}+\frac{1}{2!}\ue89e\frac{{\partial}^{2}\ue89e{F}_{3}}{\partial \xi \ue89e\text{\hspace{1em}}\ue89e\partial B}\ue89e{{\uf603}_{{\xi}_{0}\ue89e{B}_{0}}\ue89e\left(\xi {\xi}_{0}\right)\ue89e\left(B{B}_{0}\right)+\frac{1}{2!}\ue89e\frac{{\partial}^{2}\ue89e{F}_{3}}{\partial \eta \ue89e\text{\hspace{1em}}\ue89e\partial B}\uf604}_{{\eta}_{0}\ue89e{B}_{0}}\ue89e\left(\eta {\eta}_{0}\right)\ue89e\left(B{B}_{0}\right)+\dots & \left(12\right)\end{array}$  It should be noted that neither linear form of this Taylor series can be singular, by definition, and, therefore, neither is of interest. In relation to quadratic terms, coordinate substitution will be provided so that after this substitution, some free coefficients will be received that permit the zeroing of mixed quadratic terms (this approach is completely within the framework of Thom's lemma proof). On the other hand, the quadratic term (B−B_{0})^{2 }is a Morse term. Therefore, it is demonstrated that there are no singular Bdependent terms. In summary, a luminance physical coordinate does not introduce new singularities, and, because g depends only on geometrical coordinates (ξ, η) (belonging to 3D space manifold), all previous results of the WhitneyThomArnold theory apply in this new geometrical/physical (“geophysical”) 4D space. (See Table 1.)
TABLE 1 Type of Name of singularity singularity Formula Singularity Applications 0 Regular u = ξ Areas located on still image plane. v = η 1 Fold u = ξ^{2} Lines located on still image planes v = η (contours of objects). 2 Cusp u = ξ^{3 }+ ξη Points located on still image plane (Tuck) v = η (with transitions to folds and regular point). Critical points necessary for recognizing images (such as corner of mouth, eyes, etc.). 314 Other Not pre Special directions of mappings for sented here still images. Points of movie due to com frames necessary for recognizing a plexity and motion (e.g., twohumped camel lack of rotation). space. See Ref [3].  In order to explain these new results, a simple example of a homogeneous object with constant luminance B is considered.
 In such a case, Eq. (10C) does not contain Bdependence; i.e., it can be written in the following form:
 w=F _{3}(ξ, η) (13)
 Now, the first two equations (10A) and (10B) can be used without changes, to introduce (u, v)—coordinates, in the form:
 w=F(u(ξ, η), v(ξ, η) (14)
 where F is some new function, and
 w=F _{3}(ξ, η). (15)
 It is clear that wcoordinate should have the same singularities as u and v (see Table 1).
 In this case, all changes in color (gray level) will be determined only by mapping F and the contour of an object. Of course, the singularities for color w will be located at the same points as singularities of u and v. As a consequence, the singularities of color will be displaced at the contour of the object.
 Mapping of a cylinder with a given constant luminance dependence is shown in FIG. 5 and described as follows:
 B=f(ξ, η). (16)
 In a cylindrical coordinate system (where axis y coincides with the axis of the cylinder), x=α, where α is angle ∠BOA, and z is distance OB (or, radius).
 Two parametric coordinates, ξ=α, where α is angle ∠BOA (A is the central point of cylinder, B is a given point); y is the axial coordinate, and z (=R, where R is const) is the radius vector (OB). That the wparameter must be proportional to B, everywhere must be taken into account. This means that B does not create any singularities. For new coordinates on the image plane:
 u=R sin (ξ) (17A)
 v=η (17B)
 w=C·B+f(ξ, η); C=const≠0. (17C)
 On the other hand, a geometrical analysis of transformation Eq. (9A) shows that y does not produce any singularities (since y is an axial coordinate). Therefore, it can be assumed, without loss of generality, that w depends only on x and B in the following form:
 w=C·B+f(u(ξ) (18)
 or
 w=C·B+f′(x) (19)
 where f′(x)=f(u(x)).
 For critical point estimation, the Jacobian is considered, transforming coordinates (ξ, η, B) into (u, v, w) in the form:
$\begin{array}{cc}{H}_{\mathrm{ij}}=\left[\begin{array}{ccc}\frac{\partial u}{\partial x}& \frac{\partial u}{\partial y}& \frac{\partial u}{\partial z}\\ \frac{\partial v}{\partial x}& \frac{\partial v}{\partial y}& \frac{\partial v}{\partial z}\\ \frac{\partial w}{\partial x}& \frac{\partial w}{\partial y}& \frac{\partial w}{\partial z}\end{array}\right]=\left[\begin{array}{ccc}R\xb7\mathrm{cos}\ue8a0\left(x\right)& 0& 0\\ 0& 1& 0\\ C\ue89e\frac{\partial B}{\partial u}\ue89e\frac{\partial u}{\partial x}+\frac{\partial f}{\partial u}\ue89e\frac{\partial u}{\partial x}& 0& C\end{array}\right]\ue89e\text{}\ue89e\mathrm{or}\ue89e\text{}\ue89e{h}_{\mathrm{ij}}=\left[\begin{array}{ccc}R\xb7\mathrm{cos}\ue89e\text{\hspace{1em}}\ue89ex& 0& 0\\ 0& 1& 0\\ \left(C\ue89e\frac{\partial B}{\partial u}+\frac{\partial f}{\partial u}\right)\ue89eR\xb7\mathrm{cos}\ue8a0\left(x\right)& 0& C\end{array}\right]& \left(20\right)\end{array}$ 
 Therefore, it is possible to use a smooth transformation between (x, y, z) and (u, V, w).
 (ξ, B)→(u, w) (21)


 we cannot perform smooth variable substitution, and singularities exist in these points (projection of fold).




 Because the singularity appears as a result of geometrical mapping (not connected to changes of color), and we assume that the color of an object is a smooth function of coordinates x_{1}, x_{2}, it is possible to use a canonical function for representation of function f′:
 ƒ′(x _{1} , x _{3})=Bx _{3} +c·F(x _{1})
 where F is the deformation^{[3] }of a canonical polynomial (fold or x^{½} type dependence).
 The calculations presented above do not use rigorous mathematical calculations, but they are very close to Lev Landau's approach, applied successfully to many areas of theoretical physics. In his approach, the art of throwing away “inessential” terms of the Taylor series, and preserving smaller size, yet “physically important” terms, has been rigorously proven through the course of the catastrophe theory^{[3]}.
 Describing an arbitrary function by using a standard transform, such as Fourier or wavelet, is natural for periodic signal analysis. In image processing, however, these approaches have difficulties with describing very high redundancy regions with flat, slowchanging parts, as well as regions of abrupt change (or “soft edges”). Such classical description is unnatural for these types of objects because it creates excessively high input values in almost every coefficient of the Fourier transform as well as large coefficients in the case of the wavelet transform.
 At the same time, starting from Leibniz, Huygens, and Newton, a clear geometrical (polynomial) approach was developed for an analysis of smooth curves and surfaces. As discovered recently, this approach has become strongly related to many major areas of mathematics, including group theory, regular polyhedrons, wave front propagation (caustics), and dynamic systems analysis. For a clear demonstration of the unique properties of this approach, consider the classic evolvent problem, formulated in Newton's time:
 For example, for f=x^{3}, the evolvent presented in FIG. 6 can be constructed. Arnold^{[6]} has shown that the evolvent is directly related to the H_{3 }group generated by reflections of an icosahedron. (H_{3 }is a group of symmetry of the icosahedron.) H_{3 }has special properties, as described below.
 If complex space C^{3 }instead of R^{3 }is analyzed, the factorspace of C^{3 }for this group will be isomorphic to C^{3}. This means there exist some basic polynomial invariants. By using these invariants, any polynomials of this group can be represented (Arnold^{[6]}). To illustrate this property in 2D, let us describe a simplified example of three mirrors on R^{2 }as seen in FIGS. 7A and 7B.
 The points of a plane that have an equal number of reflections (12 in FIG. 4A) belong to one (regular) orbit. Points located on the mirrors belong to another orbit. A set of all irregular orbits in a factor space is a discriminant (i.e., the manifold in a factor space).
 Now the plane in 3D space can be represented, as in FIG. 4, as a plane with coordinates z_{1}, z_{2}, z_{3}. The plane can be determined by:
 z _{1} +z _{2} +z _{3}=0 (23)
 In this space, it is possible to introduce permutation of the axis, generated by reflections.
 z _{i} =z _{j} (24)
 Orbits in this context constitute a set of numbers {z_{1}, z_{2}, z_{3}} (with all permutations generated by reflections), with the additional condition of Eq. (23).
 This unordered set will be uniquely determined by polynomials:
 z ^{3}+λ_{1} z ^{2}+λ_{2} z+λ _{3}=0 (25)
 By using Eq. (25), the following is obtained:
 λ_{1}=0
 or
 z ^{3} +az+b=0 (26)
 The space of the orbits of this group will be naturally presented by the roots of a cubic polynomial Eq. (26). This means that in factor space, this space is just a plane with coordinating (a, b).
 Each point (a, b) of this space corresponds to a cubic polynomial and its roots. If some of the roots are equal, that means we have received irregular orbits.
 The discriminant in this case is
 4a ^{3}+27b ^{2}=0 (27)
 which is of^{3}/_{2 }type curve. This curve corresponds to specific orbits (in the mirrors) in FIG. 4A.
 For all other types of groups generated by reflection, analogical construction of the discriminant exists.
 It can be proven (Arnold; see Ref. [6]) that the surface creating the evolvent is diffeomorphical to the discriminant of the H_{3 }icosahedron group. As a result, by using the group representation, the redundancy of a mathematical object that is diffeomorphical to our mapping procedure has been greatly reduced.
 Taking into account the symmetry of 3D objects mapping (symmetry in a general sense, this means the Lie group, in this case) can optimally reduce redundancy and extract the information that describes the most important features of our object.
 In summary, the polynomial representation of geometrical objects (starting from: Newton through Bernoulli, up to Thom and Arnold) seems to be more natural than the common Fourier (and wavelet) approach, because polynomials are connected to groups of symmetry that permit reduction in orbit redundancy in a most natural way.
 Because the most critical part of an object—its 3D boundary—can be described by a 1D contour and three or four natural digits or “coefficients” that characterize a simple catastrophic polynomial, tremendous lossless compression of object boundaries can be achieved, far exceeding state of the art compression ratios while still preserving high quality image. Since in all state of the art still image compression methods the major information loss is at the boundaries, applying ISMP compression which actually preserves the boundary or edge information, provides unparalleled fundamental compression ratio/PSNR trade off.
 “Catastrophe” or alternatively isomorphic singular manifold as used in this patent designates a mathematical object that describes the shape of 3D object boundaries in polynomial form. The use of “catastrophe” theory for compression makes the present invention unlike all other compression methods because it helps to transmit information about 3D object boundaries without loss, preserving the features of the object most valuable to human cognition, but with very high compression rate. By applying the present invention to still image compression, a 300:1 still image compression ratio with practically invisible artifacts (PSNR equal 32 dB) and a 4,000:1 full motion image compression ratio with fully developed natural motion and good image quality, is obtained.
 The still image compression and related video compression technique of the present invention is extremely beneficial because, unlike other state of the art compression techniques, major information losses do not occur from compression at 3D object boundaries (edges) that require both high dynamic range and high resolution (i.e., both high spatial and high vertical: “Lebesque” resolution). In these edges there is a vast amount of information necessary for many processing operations vital to a quality image and human cognition. The compression technique of the present invention, unlike other compression methods, preserves intact all “soft”edge information without data loss. Hence, the present inventors have coined the term losslessontheedges (LOTE) compression. LOTE compression is possible because of the fully isomorphic projection between the 3D object boundary vicinity and its 2D projection on the screen. This fully isomorphic projection between the 3D object boundary and the 2D projection is based on Arnold's so called “catastrophe” theory that has been adapted to still image compression here. The methodology of the present invention works especially well with objects that are closer to sculptures and objects that have mostly flat surfaces combined with edgy features, i.e., very low or high spatial frequencies. This is exactly the opposite of Fourier analysis which does not work well with very low or very high frequencies. For low frequency, Fourier analysis is unsatisfactory because the coefficients must be very well balanced and at the same time can easily be hidden in noise. For high frequency components of objects such as edges, many high frequency components exist which Fourier methods eliminate. These high frequency components are what make up all important edges of the object and eliminating them reduces human cognition. ISMP analysis on the other hand, does not have this problem because it characterizes edges and objects using manifolds and hence preserves the information that makes up those edges and that was eliminated in Fourierbased compression methods.
 An understanding of how humans recognize objects will make manifest the advantage of preserving information. The retina of the human eye contains millions of receptor cells, arranged in a mosaiclike pattern in the retinal layer. The receptor cells are cones and rods. These cones and rods provide the starting point for two separate but interacting visual systems within the human eye. Cone vision is specialized for high acuity and for perception of color. Rod vision is specialized for sensitivity and the ability to distinguish color (i.e., a person can make out the general shape of the objects, but not their colors or small details^{[8]}).
 The main purpose of human vision is not to detect simple presence or absence of light, but rather to detect and identify objects. Objects are defined principally by their contours. The visual system registers a greater difference in brightness between adjacent visual images, faithfully recording the actual physical difference in light.
 David Hubel and Thornton Wiesel (Nobel prize winners in 1981) recorded the electrical activity of individual neurons in the visual cortex. They found that these cells were highly sensitive to contours, responding best not to circular spots but rather to light or dark bars or edges. They classified these cells by using a complex hierarchical system, based on their different response characteristics. In this research, the authors outlined that the perception of long and linear bars provided maximum response in the human visual system.
 Human brain zones, which decode specific properties of image recognition, are spatially organized in the brain according to their function. Thus, different localized sets of neurons in the visual cortex are specialized to carry codes for contours, color, spatial position and movement. This segregation of functions explains why a person who has had a stroke, which damaged part of the cortex, sometimes loses the ability to see contours without losing the ability to see colors.
 Special mechanisms of object edge extraction in the human visual system allow extraction of important objects from a background, even if the object has bulk colors very close to the colors of the second plane. The latter feature is extremely important for registration of military targets, and makes ISMP an effective compression algorithm for ATR.
 This still image compression performance can be transformed into analogous video image compression through the typical 10:1 factor for state of the art video image compression. Therefore, the inventive technique can be applied not only to high resolution digital video/still image transmission, but also to multimedia presentation, high quality video conferencing, video servers, and the storage of large amounts of video information.
 Video compression is a four dimensional (4D) problem where the goal is to remove spatial and temporal redundancy from the stream of video information. In video there are scenes containing an object that continuously changes without jumps and has no edges, and, on the other hand, there are also scenes where there are cuts which are big jumps in the temporal domain or big jumps in the spatial domain (such as “edges”). These abrupt changes or jumps can be described as “catastrophes.” Using catastrophe theory, these behaviors can be described by one or more elemental catastrophes. Each of these elemental catastrophes describes a particular type of abrupt change in the temporal or spatial domains. In general, categorizing catastrophes in 4D space is even less established than catastrophe general which is relatively unknown. Furthermore, 4D space is far less understood than 3D space but, similarities between them can be expected and can use projectiontype mapping can be used, but in temporal space. One solution is to use spatial catastrophes along with temporal catastrophes.
 In order to apply catastrophic theory to video imagery, a fourth “geometrical” coordinate, time leading to timespace (4D) is preferably added. In the case of the inventive isomorphic singular manifold projection (ISMP) methodology, fivedimensional (5D) geometrophysical space (x, y, z, t, B), where B is brightness, or luminance, is obtained. This 4D timespace (x, y, z, t) plus physical coordinate, B, can be split into 4D geometrophysical space, and time (t) and treated separately except in the case of relativistic velocities. In the latter, relativistic case, the 5D space can be analyzed by Poincare group formalism. In the common, nonrelativistic case, however, temporal singularities (catastrophic) may be described in the timeluminance (t, B) domain only. The timeluminance singularities may interfere with spatial singularities (previously discussed). In such a mode of operation, each block of the image is represented by a single timevariable value.
 According to FIG. 8, there are only two possible singularities describing any type of mapping including smooth curve projection designated (1) shown in FIG. 8 where <B> is the average Bvalue characterizing a frame as a total structure (the smooth projection shown in FIG. 8 represents movement of a physical object. Item (2) in FIG. 8 representing a catastrophic frame change, and item (3) representing position/tilt/zoom camera changes. The critical <B>parameter may be, for example, an average blocktoblock error (e.g. mean square error). In summary, temporal catastrophic formalisms can be applied to MPEG hypercompression by replacing the average error parameters by integrated luminancechanges.
 One way to represent these 4D catastrophes is to use well known 3D projections or mapping catastrophes which were discovered in the early 1980's. These “transformations” or “reconstructions” or “metamorphoses” in time are 4D problems which can be separated into two 3D problems: 1) Spatial catastrophes may be defmed in 3D space (x, y, B) such as occurs when there is a large change in intensity, B, over a small change in x, y; 2). Temporal catastrophes may also be defined temporally such as occurs where there is an abrupt change in motion over time such as is present during the rotation of an object or a cut from one scene to another. The 3D temporal problem can be further reduced to a 2D problem by transferring the (x, y, B) coordinates into 1D merit space. Merit space is defmed by the lack of similarity between frames in time.
 Images are 3D distributions of intensity. Abrupt changes in intensity occurring over small changes in x, y may be treated as catastrophic changes. The inventors have modified catastrophe theory to fit images and to solve the problems of image and video compression. The inventors have introduced a physical coordinate, B (luminance) into conventional geometrical coordinates to create “geometrophysical” surfaces.
 The exists a finite list of fourteen polynomials or “germs” which describe different edge transitions or projections in mapping in 3D space. Typically, only about three of these polynomials or germs are necessary to describe virtually every edge effect. The others are used on occasion to describe spatial projections.
 The germs of the projections are equivalent to the germs of the projections of the surfaces z=f(x,y) along the xaxis. The table below identifies the fourteen polynomials of germs.
Type f(x, y) Type f(x, y) 1 x (without singularities) 8 x^{4 }+ x^{2}y + xy^{2} 2 x^{2 }(fold) 9, 10 x^{5 }± x^{3}y + xy 3 x^{3 }+ xy (Whitney's tuck) 11, 12 x^{3 }± xy^{4} 4, 5 x^{3 }± xy^{2 }(3/1 type curve) 13 x^{4 }+ x^{2}y + xy^{3} 6 x^{3 }+ xy^{3 }(9/2 type curve) 14 x^{5 }+ xy 7 x^{4 }+ xy (4/3 type curve)  In theory, a projection of a surface does not have any germs that are inequivalent to the fourteen germs in the above table. It should be understood that the Spectral Series for Reduction to the Normal Form (SSRNF) method is used for the unique reduction of the arbitrary polynomial to the germs presented in the above table. It is presented here only in a descriptive form:
 Let e_{1}, . . . , e_{n}quasihomogeneous polynomial (N+p degree) that generates
 A _{p} ^{τ+1}
 for diffeomorphism.

 and that the series f=f_{0}+f_{1}+ . . . after substitution has a form
 ƒ(y _{1} , . . . , y _{n})=ƒ_{0}(x)+ƒ_{1}(x)+ . . . +ƒ_{p−1} +Σc _{i} e _{i}(x)+R, R εA _{p+1}
 and c_{i }represent the numbers.
 Catastrophe theory has not been before used for studying image intensity because the number of coefficients necessary to satisfactorily describe an image using standard polynomials is simply too large and can exceed the number of pixels present in an image. Obviously, such an analysis is not worthwhile because the data that need to be handled are larger than the number of pixels, itself a very large number. The inventors have discovered that it is possible to remove many of the details or “texture” in images, leaving the important “sculpture” of the image, prior to characterizing the image with polynomials, to significantly decrease the number of coefficients in the polynomials that describe the different edge transitions in mapping and 3D space.
 The following is an abbreviated description of the still image compression method as in the flow chart of FIG. 9. Step1 involves segmenting the original image into blocks of pixels, for example 16×16. Step 2 is to create a model surface for each segment or block corresponding to the original image so that there is isomorphism between the original image and the polynomial surface in accordance with Arnold's Theorem. More particularly this may involve calculating the equation F_{modelled }for each block or segment by substituting for variables in canonical polynomials. (See Steps 37 of detailed flow chart which follows). This step inherently eliminates texture of the image and emphasizes the “sculpture” characteristics. Step 3 is to optimize each model segment. This is done by calculating the difference between the original and model segments and choosing coefficients for the canonical polynomial which have the lowest Q i.e., the smallest amount of difference between the original segment and the modelled segment. This is repeated on a segment by segment basis. ( See Steps 812 of the detailed flow chart which follows). Step 4 is to find connections between adjacent segments to create an entire image i.e., a model image of the entire frame. (See Steps 1418 of the detailed flow chart). This yields an entire image that has only the “sculpture” characteristics of the original image and eliminates texture. Step 5 is to calculate the peak signal to nose ratio PSNR over the entire image and where the PSNR of the entire image is less than a threshold, the difference between the original image and the modelled image is calculated. This step recreates the texture information of the original image that was lost during the process. Thus, after this step there are two sets of data: the “sculpture” characteristics represented by a few discrete numbers or “datery” and the texture information of the image. (See Steps 1921) Step 6 is to use standard lossy compression on the texture portion of the data and then to combine the texture and datery and apply standard lossless compression to that combined data. (See Steps 2224 of the detailed flow chart).
 Now the preferred still image encoding method will be described in detail in relation to the detailed flow chart.
 In the following description of the still image encoding process according to the present invention the following definitions are used:
 I_{o}=original I frame
 I_{m}=modeled I frame
 I_{d}=difference I frame (I_{d}=I_{o}−I_{m }for each frame)
 i_{o}=segment or block of original frame
 i_{m}=segment or block of modeled frame
 i_{d}=segment or block of difference frame (i_{d}=i_{o}−i_{m }for each block)
 Referring now to FIG. 10, this figure sets forth a flow chart of the still image encoding process according to the present invention. In step1, the next still image or I_{o }frame is captured. If only still images are being compressed for still image purposes, this image will represent one of those still images. If video is being compressed, the still image to be compressed here is one of the video's I frames which will be compressed in accordance with this method and then inserted at the appropriate location into the video bitstream.
 In step2, the original image I_{o }is segmented into blocks of pixels of any desired size such as, for example, 16×16 square blocks. The original image is seen in FIG. 11A. Any segment size may be used as desired. These segments or blocks of pixels are designated i_{o}. This segmentation is done according to standard segmentation methods. As an example, the total number of segments or blocks for a 512×512 image is 512×512/(16×16)=1024 different noninterleaving 16×16 segments or blocks.
 Step3 is the first step involving segment by segment operation on each i_{o }using matrix representation of each segment. In step 3, the Dynamic Range (R) of each segment or block is calculated using the following equation:
 R=max(i_{o})−min(i_{o})
 In the above formula, the pixel having the maximum intensity is subtracted from the pixel having the minimum intensity in the segment. This difference is the Dynamic Range R.
 Step4 compares the Dynamic Range R to R_{o }which is a threshold determined from trial and error. The threshold R_{o }is chosen so as to eliminate unnecessary compression such as compression of background scenes. In this regard, if the value R is very small and less than R_{o}, the image is most likely background and the compression technique of the present invention is not needed. In this case, the process is started over again between steps 2 and 3 and another segment or block is operated on. If R is greater than or equal to R_{o }then the subsequent steps involved in choosing a canonical polynomial from the table and creating a model polynomial by solving its coefficients are then performed. This set of steps now generally described involves choosing the polynomial from the table which best matches each particular segment or block.
 Turning now to step5, a first canonical polynomial from the table is taken. In step 6, substitutions for variables in the canonical polynomials are found. It is possible to apply (1) a nonhomogeneous linear transformation (shift of coordinates), (2) a homogeneous linear transformation (rotation of axis) or (3) a nonhomogeneous nonlinear transformation. For example, if the canonical polynomial
 ƒ_{canonical} =x _{1} ^{3} +x _{1} x _{2}
 is taken from the table, variables x_{1 }and x_{2 }are substituted for as follows using the third example above nonhomogeneous nonlinear transformation:
 x _{1}=(y _{1} +a _{1} y _{1} ^{2} + . . . a _{n} y _{n} ^{2}); x _{2}=(y _{2} +b _{1} y _{2} ^{2} + . . . b _{n} y _{n} ^{2}).
 From this substitution a function describing a “modeled” surface (as opposed to the original image surface) is generated as follows:
 ƒ_{model}=(y _{1} ^{2} +a ^{2} y _{1} ^{4}+2ay _{1} ^{3})(y _{1} +ay _{1} ^{2})+y _{1} y _{2} +aby _{1} y _{2} ^{2} +ay _{1} y _{2} +by _{1} y _{2} ^{2} =y _{1} ^{3} +a ^{2} y _{1} ^{5}+2ay _{1} ^{4} +ay _{1} ^{4} +a ^{3} y _{1} ^{6}+2a ^{2} y _{1} ^{5} +y _{1} y _{2} +aby _{1} y _{2} ^{2} +ay _{1} y _{2} +by _{1} y _{2} ^{2}=(y _{1} +ay _{1} ^{2})^{3}+(y _{1} +ay _{1})(y _{2} +by _{2} ^{2})
 At step7, the modeled surface is created by substituting the coordinates of each pixel in the original segment or block into the equation f_{model}. A modelled surface is seen in FIG. 11C. This creates a matrix containing the values f_{m(1,1)}, x_{m(1,2) }. . . as seen in FIG. 12. Specifically, this matrix is created by substituting the coordinate of the pixel 1,1 from the original segment into the equation f_{model }to generate the element f_{m(1,1) }in the modeled matrix. Next, the coordinate of the pixel 1,2 from the original segment is substituted into the equation f_{model }to generate value f_{m(1,2) }which goes in the 1,2 pixel location of the modeled surface. This is done for each pixel position of the original segment to create a corresponding modeled matrix using the equation f_{model}.

 In other words, Q is calculated by subtracting corresponding pixels from the i_{o }segment (the original segment) from the i_{m }(the modeled segment) and squaring this subtraction and summing up all these squares.
 At step9, Q is compared to a predetermined threshold Q_{o }based on image quality desired. Q should be less than Q_{o }because the point of the step 8 is to minimize the sum of the differences between the analogous pixels in the original and modeled frames so as to generate a modeled surface that is as close as possible to the original surface. If Q is greater than Q_{o}, the procedure loops back up to step 6 where new coefficients are tried in the same polynomial. Then steps 7, 8, and 9 are repeated, and if Q is less than Q_{o }with that new set of coefficients, then the process continues into step 10 where that Q and the coefficients that produced the lowest Q for that polynomial are stored. After storage at step 10, the process loops back up to step 5 if all polynomials have not yet been tested where the next canonical polynomial from the library is chosen and tested and solved for coefficients which produce the lowest Q for that next polynomial. Hence, steps 6, 7, 8, and 9 are repeated for that next polynomial until coefficients are found which produce the lowest Q for that polynomial. At step 10, the Q and the coefficients for that next polynomial are stored. This process of steps 5, 6, 7, 8, 9, and 10 is repeated for each polynomial in the library. After each polynomial in the library is tested for the segment under test, the process moves to step 11.
 At step11, the polynomial having the lowest Q of the polynomials tested for that segment is chosen. That polynomial is transferred to step 12.
 At step12, all coefficients for the chosen polynomial (the one having the lowest Q of all the polynomials tested for that segment) are stored. These coefficients are coefficients of the equation f_{model }which describes the modeled surface.
 After step12, the next set of operations involves segment by segment operation working only with the polynomials and their coefficients whereas the above steps 512 worked with the matrix representation of each segment. Because only the polynomials and their coefficients are worked with in the next set of operations, a significant amount of compression has taken place because the data representing the surface is far less voluminous than when a matrix representation of the segments is worked with. The data is simply coefficients of polynomials which can be called “datery”.
 At step13, the current segment is taken or captured from the above steps. At step 14, a connection is found between adjacent or neighboring segments by extending the surface of a first segment into a second segment and finding differences between the extended surface and the second segment surface. Specifically, this is done by finding the average distance “q” between the surface which extends from the first segment into the second segment and the surface of the second segment using standard methods. If the average distance “q” is smaller than a threshold value q_{o}, the surface of the second segment is approximated by the extended surface. In other words, if the distance q is smaller than the threshold value q_{o}, the second segment surface is thrown out because it can be approximated satisfactorily by substituting the extended surface in its place. If the average distance q is greater than the threshold value q_{o}, a connection needs to be found between the extended surface and the surface of the second segment.
 Thus, at step15, the average distance q is checked to determine whether it is less than the threshold value q_{o}. If it is, then the connections between the adjacent or neighboring segments which can be plotted as a graph, as seen in FIG. 13, are stored on a segment by segment basis. In other words, as seen in FIG. 13, the surfaces which extend from, say, a surface in segment “9” into adjacent or neighboring segments (8, 10, 14 and 15), if any, are stored in the polynomial for segment 9 (earlier calculated and then stored at step 12) which then represents that graph of connections between segment 9 and segment 8, 10, 14 and 15. In other words, the polynomial that was calculated and stored for the segment in question, here segment 9, is modified so that it now extends into adjacent segments 8, 10, 14, and 15 and represents the surfaces in those segments. The polynomials for segments 8, 9, 10, 14 will be substituted with the new bigger scale polynomial obtained from 9.
 If the average distance from9 was not less than q_{o }(which indicates that the surface extended from the segment in question, segment 9 for example, into an adjacent segment, 8, 10, 14, or 15 for example, did not satisfactorily approximate the surface of the second segment), then a spline must be calculated at step 17.
 At step17 splines are calculated from the segment with adjacent segments using standard spline equations which need not be detailed here.
 After both steps16 and 17, the process continues at step 18. At step 18, a model image i_{m }is created of the entire frame by creating a table of all segments for that frame using the information calculated for each segment in the above steps. The creation of this table representing the entire frame from its numerous segments is analogous to step 7 where a modeled segment was created by substituting the pixel coordinates from the original segment into the f_{model }polynomial to get a matrix describing the modeled surface. At step 18, however, instead of creating a modeled segment of pixels, a modeled frame is created from modeled segments. Thus, it can be seen that the smaller parts calculated above are now being combined to generate an entire modeled frame.
 After step19, the peak signal to noise ratio (PSNR) is calculated over the entire image using the equation:
$\mathrm{PSNR}=\frac{10\ue89e\text{\hspace{1em}}\ue89e{{\mathrm{log}}_{10}\ue8a0\left(\mathrm{number}\ue89e\text{\hspace{1em}}\ue89e\mathrm{of}\ue89e\text{\hspace{1em}}\ue89e\mathrm{gray}\ue89e\text{\hspace{1em}}\ue89e\mathrm{scale}\ue89e\text{\hspace{1em}}\ue89e\mathrm{levels}\right)}^{2}}{Q}$ $Q=\sqrt{\frac{1}{\mathrm{hv}}\ue89e\sum _{x=0}^{M1}\ue89e\text{\hspace{1em}}\ue89e\sum _{y=0}^{N1}\ue89e{\left[{I}_{o}\ue8a0\left(x,y\right){I}_{m}\ue8a0\left(x,y\right)\right]}^{2}}$  h and v are number of pixels in horizontal and vertical directions respectively for the entire frame image. The Q values for each of the segments were stored at step10 above and may be retrieved for this purpose.
 At step20, the PSNR of the entire frame is compared to a threshold P_{o}. If PSNR is less than P_{o}, then no further processing according to the present invention need be accomplished and processing can continue at step 24 where a standard lossless compression such as Hoffman encoding and runlength encoding are used to further compress the frame data. The compressed data is then sent to storage or a communication link.
 If PSNR is greater than P_{o }at step 20, then processing continues at step 21. At step 21, the difference between the original frame I_{o }and the modeled frame I_{m }is found by subtracting each pixel in I_{m }(which was created at step 18) from the corresponding pixels in I_{o }and a new frame I_{d }is created (see FIG. 14) where each pixel in that frame has as its value the difference between the corresponding pixels in the frame I_{o }and the frame I_{m}. The frame I_{d }therefore corresponds to the high frequency components, such as edge information which typically is lost in conventional compression techniques. This “texture” information containing high frequency components and edge information is then compressed separately at step 22.
 At step22, standard lossy texture compression of the newly created frame I_{d }is performed by using standard methods such as DCT, wavelet, and fractal methods. At step 22, standard additional lossless compression is also performed. The output of step 22 is I_{d}′ which then is fed into step 23. At step 23, the I_{m }frame is stored and the I_{d}′ frame is stored. This concludes the compression of the still frame or I_{o }frame.
 As can be seen, the polynomial surface image is highly compressed because it is stored and transmitted as a complex algorithm (polynomial) rather than as a matrix representation. Additionally, the edge contour image I_{d }is separated from the polynomial surface, as a byproduct of characterizing the original image by a canonical polynomial and contains the high frequency and edge components and is itself compressed.
 The still image decoding process will now be described as seen in the flow chart of FIG. 15. The input to the still image decoding process will be either just the whole frame I_{m }in the case where the PSNR of the whole frame at step 20 was not less than threshold P_{o }or the whole frame I_{m }plus I_{d}′ where the PSNR of the whole frame at step 20 was less than P_{o }and the differences between the original I_{o }frame and the modeled frame I_{m }were calculated to create new frame I_{d }holding the textured or high frequency and edge information.
 In either case, the first step in decoding is step1 which decodes the lossless compression data from the encoder which was compressed at step 24. At step 2 of the decoding process, frame I_{m }is separated from the other data in the bitstream. At step 3, the first graph or segment which was stored at step 16 on a segment by segment basis is taken.
 At step4, whether the segment belongs to a graph (i.e., has connections to adjacent segments) or is an isolated segment (i.e., has no connections to neighboring segments) is tested. If the segment does belong to a graph, then at step 5, a segment i_{m }is constructed for each graph (analogous to the creation of the modeled matrix surface in step 7 of the encoding process) using the polynomial that was stored for that graph at step 16 of the encoding process.
 If the segment does not belong to a graph, then after step4 the process skips step 5 and continues with step 6.
 At step6, the separate graphs using standard splines are connected. In other words, those segments from steps 14, 15, and 17 which were connected by splines will be reconnected here. (Recall that it was these segments for which the extended surface of another adjacent segment did not satisfactorily characterize the surface of these segments and therefore a spline equation had to be used.)
 From step5 where a segment i_{m }for each graph was reconstructed, and from step 6 where separate graphs were connected using standard splines, the process continues at step 7.
 At step7, the frame I_{m }is constructed using segments from step 5 (in the same way as the frame I_{m }was constructed in the encoding process at step 18 and also similar to how an individual segment or modeled surface or modeled segment was created at step 7 in the encoding process.)
 At step8, the presence of a I_{d}′ frame for or in conjunction with the frame I_{m }is tested. If there is no frame I_{d}′, then the process is finished and the still image is fully decoded for that frame. If on the other hand there is a frame I_{d}′ in conjunction with the frame i_{m}, then the process continues to step 9.
 At step9, the frame I_{d}′ is decompressed.
 At step10, frame I_{o}′ is created from the combination of frame I_{m }from step 7 of the decoding process and frame I_{d}′ from step 9 of the decoding process. After the frame I_{o}′ is created, the process is finished and the still image is fully decoded.
 The inventive compression technique for still images can be incorporated into standard MPEG compression to enhance video compression through spatial hypercompression of each I frame inserted into the video bitstream. Alternatively, in the preferred embodiment, a novel motion estimation technique is employed which provides significantly greater compression due to temporal compression. According to the present invention, I frames are inserted according to video content. This is done by accumulating the error or difference between all corresponding microblocks or segments of the current frame and the predicted frame and comparing that accumulated error or difference to a threshold to determine whether the next subsequent frame sent should be an I frame. If the error or difference is large (i.e., when motion error is high), the I frame is sent. If the error or difference is small, the I frame is not sent and the frame sequence is unaltered. As a consequence, full synchronization of I frame insertion with changes in scene is achieved and bandwidth is significantly reduced because I frames are inserted only where necessary, i.e., where content requires them. Thus, the present invention, for the first time, analyzes the errors between the I frame and the B and P frames into which it will be inserted to decide whether to insert the I frame at that point or not. Consequently, the present invention significantly increases the overall image compression ratio, while offering a simultaneous benefit of increased image quality. In addition, by using the technique of the present invention for video compression, the distances between I frames are enlarged, which leads to better motion estimation and prediction.
 The video compression technique of the present invention may be used with both I frames compressed using the still ISMP compression encoding process of the present invention or standard I frame compression techniques. The most significant compression will occur if both the ISMP compression encoding process of the present invention and the motion estimation process of the present invention are used. It is worth noting that in existing systems, a reasonable quality video can be produced only if I frame compression is not higher than 20:1 to 40:1. With the present invention I frame compression of 300:1 is achieved. The following table illustrates the improvement over standard compression of the inventive technique of fixed separation of I frames compressed with the inventive CT algorithm used in conjunction with the inventive variable separation of I frames compressed with the CT algorithm.
Fixed Variable separation separation of I with I frames frames compressed compressed with CT with CT Standard Compression algorithm algorithm Image 352 × 240 352 × 240 352 × 240 Resolution (8bit per pixel) Uncom 84,480 84,480 84,480 pressed Im age Size per Frame I Frame 30:1 300:1 100:1 Compression Compressed 2816 250 860 Image Size per Frame I Frame 15 frames 15 frames 45 frames Separation (0.5 second) (0.5 second) (1.5 second) Average Size 422 422 422 of BP Frame (200:1) Uncom 84,480*30 freq*60 sec = 1,520,064,000 1,520,064,000 pressed Data 1,520,064,000 Size for 1 min. Video Overall 10,483,000 1,900,080 1,520,000 Compressed Data Size for 1 min. Video Correspond 145:1 800:1 1000:1 ing Com pression Ratio  Motion estimation is important to compression because many frames in full motion video are temporally correlated, e.g., a moving object on a solid background such as an image of a moving car will have high similarity from frame to frame. Efficient compression can be achieved if each component or block of the current frame to be encoded is represented by its difference with the most similar component, called the predictor, in the previous frame and by a vector expressing the relative position of the two blocks from the current frame to the predicted frame. The original block can be reconstructed from the difference, the motion vector, and the previous frame. The frame to be compensated can be partitioned into microblocks which are processed individually. In a current frame, microblocks of pixels, for example 8×8, are selected and the search for the closest match in the previous frame is performed. As a criterion of the best match, the mean absolute error is the most often used because of the good trade off between complexity and efficiency. The search for a match in the previous frame is performed in a, for example, 16×16 pixels window for an 8×8 reference or microblock. A total of, for example, 81 candidate blocks may be compared for the closest match. Larger search windows are possible using larger blocks 8×32 or 16×16 where the search window is 15 pixels larger in each direction leading to 256 candidate blocks and as many motion vectors to be compared for the closest match.
 Once the third subsequent frame is predicted, the standard methods provide that the error between a microblock in the current frame and the corresponding microblock in the predicted frame are compared and the error or difference between them is determined. This is done on a microblock by microblock basis until all microblocks in the current frame are compared to all the microblocks in the predicted frame. In the standard process these differences are sent to the decoder real time to be used by the decoder to reconstruct the original block from the difference, the motion vector, and the previous frame. The error information is not used in any other way.
 In contrast, in the present invention, the error or difference calculated between microblocks in the current frame and the predicted frame are accumulated or stored and each time an error is calculated between a microblock in the current frame and the corresponding microblock in the predicted frame that error is accumulated to the existing error for that frame. Once all the errors for all the blocks in the current frame as compared to the predicted frame are generated and summed, that accumulated error is then used to determine whether a new I frame should be inserted. This methodology is MPEG compatible and yields extremely high quality video images not possible with state of the art motion estimators. The accumulated error is used to advantage by comparing it to a threshold E_{0 }which is preset depending upon the content or type of the video such as action, documentary, or nature. If E_{0 }for a particular current frame is exceeded by the accumulated error, this means that there is a significant change in the scene which warrants sending an entire new I frame. Consequently, an entire new frame is compressed and sent, and the motion estimation sequence begins again with that new I frame. If E_{0 }is not exceeded by the accumulated error, then the differences between the current frame and the predicted frame are sent as usual and this process continues until E_{0 }is exceeded and the motion estimation sequence is begun again with the sending of a new I frame.
 Now turning to FIG. 16, the motion estimation process is now described in detail. At step1, the next F_{0 }frame is taken. This frame may be the first frame of the video in which case it is an I frame or may be a subsequent frame. At step 2, if F_{0 }was compressed by standard DCT methods, the left branch of the flow chart of FIG. 8 is followed. If F_{0 }was compressed using the inventive ISMP algorithm, the right branch of the flow chart in FIG. 8 is followed.
 First, assuming that F_{0 }was compressed using standard DCT methods, step 3 involves standard segmenting of the F_{0 }frame into search blocks having subblocks called microblocks and defining motion vectors which are used to predict the third subsequent frame after F_{0}. This is accomplished using standard techniques well known in the art.
 At step4, the error or difference between each microblock in F_{0 }and the corresponding microblock in the predicted third subsequent frame is defined for all microblocks in F_{0}. At this point the inventive motion estimation processed diverges from standard techniques.
 If a standard MPEG encoderdecoder scheme was being used, these microblock differences would be sent from the encoder to the decoder and used by the decoder to reconstruct F_{0}. By sending only the differences between F_{0 }and the predicted third subsequent frame, significant compression is realized because it is no longer necessary to send an entire frame of information but only the differences between them. In accordance with standard MPEG encoderdecoder techniques, however, a new I frame is necessarily transmitted every 15 frames whether an I frame is needed or not. This poses two problems. Where the I frame is not needed, bandwidth is wasted because unnecessary bits are sent from the encoder to the decoder (or stored on disc if the process is not done real time). On the other hand, where the content of the video is such that significant scene changes occur from one frame to another much more often than every 15 frames, the insertion of an I frame every 15 frames will be insufficient to ensure a high quality video image at the decoder. For these reasons, the motion estimation technique of the present invention is especially valuable because it will, dependent upon the content of the video, insert or send an I frame to the decoder when the content of the video warrants it. In this way, a high quality image is maintained.
 This is accomplished in the present invention by, as seen at step5, accumulating the error between corresponding microblocks in the F_{0 }and the predicted third subsequent frame as each error is defmed in step 4 for each microblock of F_{0}.
 The next step, step6, is optional and involves normalizing the total accumulated error for F_{0 }by defining an average error A which is the total accumulated error divided by the number of microblocks in F_{0}. This yields a smaller dynamic range for the errors, i.e., smaller numbers may represent the errors.
 Continuing with step7, the accumulated error (whether normalized or not) is compared to a threshold error E_{0}. E_{0 }is chosen based upon video content such as whether the video is an action film, a documentary, a nature film, or other. Action videos tend to require insertion of I frames more often because there are more drastic changes in scene from one frame to another. It is especially important when compressing such videos to use the motion estimation technique of the present invention which can insert additional I frames based on video content where necessary to keep video image quality high. In choosing E_{0}, bandwidth versus quality should be considered. If E_{0 }is set high, a high level of errors will be tolerated and fewer I frames will need to be inserted. Quality, however, will decrease because there will be an under utilization of bandwidth. If, on the other hand, E_{0 }is set too low, I frames will be inserted more frequently and available bandwidth may be exceeded and frames may start to drop out as commonly happens with MPEG. So the threshold E_{0 }should be tuned to video content. This can be done in real time by analyzing the video offline and varying E_{0 }in accordance with the statistics of the video, such as the number of cuts, the amount of action, etc. This process may be enhanced by using genetic algorithms and fuzzy logic. Where the accumulated error is greater than E_{0}, the next frame sent will be an I frame. In accordance with standard techniques, it is preferable that the I frame be compressed prior to sending it to the decoder. This reinitiates the sequence of frames at step 8.
 If the accumulated error is less than E_{0}, the subsequent frame is not sent as an I frame but the differences are continued to be sent at step 9 to minimize bandwidth of the signal, sent between the encoder and decoder. The process then reinitiates at step 1 where the next frame F_{1 }is taken. That next frame may not be an I frame but may instead be a subsequent frame, and the methodology is the same in either case. The next frame, whether it is an I, B, or P frame, is compared to the predicted third subsequent frame and the method continues as described above.
 In an alternative embodiment, instead of sending the I frame as the next subsequent frame, the I frame could be sent as the current frame and used to replace error data for each microblock data for each microblock stored in the decoder buffers. This could be accomplished by clearing the buffers in the decoder holding errors between each of the microblocks F_{0 }and the predicted third subsequent frame and replacing that data with the I frame. Although not compatible with MPEG, it may be advantageous in certain situations to clear out the buffers containing the high error frame data and replace that data with the next frame as an I frame.
 The motion estimation technique of the present invention may also be used to dynamically change or update compression ratio on a frame by frame basis by providing feedback from the receiver or decoder and using that feedback to change parameters of the compression engine in the registers of the video compression chips. For example, if the accumulated error calculated in the motion estimation technique of the present invention were too frequent or extraordinarily high, this information could be used to alter the parameters of the compression engine in the video compression chips to decrease the compression ratio and thereby increase bandwidth. Conversely, if the accumulated error over time was found to be unusually low, the compression ratio could be increased and thereby the bandwidth of the signal to be stored could be decreased. This is made possible by the accumulation of errors between the corresponding microblocks of the current frame (F_{0}) and the predicted third subsequent frame. This is not possible in prior art techniques because, although the error between corresponding microblocks of the current frame and the predicted third subsequent frame are calculated, there is no accumulated error calculated and no use of that accumulated error anywhere in the system. In the present invention, however, the accumulated error is calculated and may, in fact, be used on a frame by frame basis to decide whether the next frame should be an entire I frame as opposed to only the difference signal.
 In a bandwidth on demand system, for example, if the feedback from the receiver indicates that there is a high bit error rate (BER), the transmitter may lower the bandwidth by increasing the compression ratio. This will necessarily result in a signal having sequences of different bit rates which are not possible in prior art MPEG systems. Intelligent systems such as genetic algorithms or neural networks and fuzzy logic may be used to determine the necessary change in compression ratio and bandwidth offline by analyzing the video frame by frame.
 Turning now to the right branch of FIG. 16, this branch is followed if the still compression method selected was the ISMP algorithm of the present invention which compresses each frame in accordance with catastrophic theory and represents the “structure” of that image in a highly compressible form using only the coefficients of canonical polynomials. Step3A in the right branch would be to predict the third subsequent frame from the current frame (here F_{0}) using standard techniques of defining the motion vectors of microblocks within the search blocks by template matching.
 Step4A would be to define the error between microblocks in F_{0 }and the microblocks in the predicted third subsequent frame. This is done using standard techniques. If a particular microblock in F_{0 }has a match with a microblock in the predicted frame, i.e., the error is 0, then the coefficients of the polynomial that were generated for that microblock when F_{0 }was compressed using the ISMP algorithm are then sent to the decoder and used along with the motion vectors generated in step 3A to reconstruct F_{0}. The sending of just the coefficients results in much higher than normal compression because the number of bits representing those coefficients is very small. If a microblock in F_{0 }has no match in the predicted third subsequent frame i.e., an error exists between those corresponding microblocks, new coefficients are generated for the corresponding microblock in the predicted third subsequent frame and those coefficients are sent to the decoder and used along with the motion vectors generated in step 3A to reconstruct F_{0}. As an alternative, the newly generated coefficients for the corresponding microblock in the predicted third subsequent frame could be subtracted from the coefficients of the corresponding microblock in F_{0 }to even further compress the data. This may be done but is not necessary because the coefficients representing each microblock constitute highly compressed data already and further compression is not necessary.
 At step5A, the errors from the above comparison of F_{0 }and P are accumulated.
 At step6A, the accumulated errors are normalized by the number of microblocks.
 At step7A, the accumulated error is compared to the threshold E_{0}. And if the accumulated error is greater than the threshold E_{0 }a new I frame is sent as the new subsequent frame at step 8A. If the accumulated error is less than the threshold error E_{0 }the coefficients that were newly generated for a particular microblock that did not find a match are continued to be sent to the decoder at step 9A. After both steps 8A and 9A the process reinitiates at step 1. Thus, according to the present invention, the error data is used and interpreted in a novel way which provides high compression and quality imaging.
 Referring now to FIG. 17, the hardware for performing motion estimation is depicted in block diagram format. All of the hardware is standard. A host computer10 communicates with video processor board 12 over PCI bus 14. The host computer 10 is preferably of at least the 100 MHz pentium class. PCI bus controller 12 controls communications over the PCI bus. EPROM 14 stores the coefficients and transfers them to the PCI bus controller 12 so that all the internal registers of the PCI bus controller 12 are set upon startup. Input video processor 16 is a standard input video processor. It is responsible for scaling and dropping pixels from a frame. It has two inputs, a standard composite NTS signal and a high resolution Y/C signal having separated luminance and chromance signals to prevent contamination. The input video processor 16 scales the normal 702×480 resolution of the NTS input to standard MPEG1 resolution of 352×240. The input video processor 16 also contains an A/D converter which converts the input signals from analog to a digital output.
 Below input video processor16 is audio input processor 18 which has as its input left and right stereo signals. The audio input processor 18 performs A/D conversion of the input signals. The output of the audio input processor 18 is input to a digital signal processor (DSP) audio compression chip 20 which is standard. The output of the audio compression chip 20 is input into the PCI bus controller 12 which can place the compressed audio onto the PCI bus 14 for communication to the host computer 10. Returning to the video side, the output of the input video processor 16 is input to an ASIC 22 (Application Specific Integrated Circuit) which is one chip of a three chip video compression processor also having a DTC based compression chip 24 and a motion estimator chip 26. The ASIC 22 handles signal transport, buffering and formatting of the video data from the input video processor 16 and also controls both the DTC based compression chip 24 and motion estimator chip 26. All of these chips are standard. An output of each of the chips 22, 24, and 26 of the video compression processor 23 is input to the PCI bus controller 12 for placing the compressed video on the PCI bus for communication to the host computer 10.
 The compressed video stream from the video compression processor23 on the board 13 undergoes lossless compression in the host computer using standard lossless compression techniques such as statistical encoding and runlength coding. After that lossless compression, the audio and video are multiplexed in standard fashion into a standard video signal. In order to have synchronization of audio and video the packets containing video and audio must be interleaved into a single bit stream with proper labeling so that upon playback they can be reassembled as is well known in the art.
 Importantly, the errors that were calculated in the motion estimator26 between the current frame and the predicted third subsequent frame are transmitted to the host computer 10 over the PCI bus 14 so they can be transmitted to the encoder (not shown) to recreate the current frame at the encoder using that error or difference signal and the motion vectors generated during motion estimation. This is standard in the art. In accordance with the motion estimation of the present invention, however, that error is also accumulated in the host computer in a software routine in accordance with the motion estimation techniques of the present invention.
 Referring now to FIG. 18 is a flow chart describing error accumulation in the motion estimation procedure. At Step1 the error buffer in the compression processor 23 is read through the PCI bus 14. At Step 2 that error is accumulated in an error buffer created in software in the host computer 10 so that the accumulated error will equal the preexisting error plus the present error. At Step 3 the accumulated error is compared to a threshold error and if the accumulated error is larger than the threshold error then a new I frame is sent and the error buffer in the compression processor need not be read again for that particular frame. If the accumulated error is not greater than the threshold error then the process loops back up to Step 4 where the next subsequent microblock in that frame is chosen. If there is a subsequent microblock in that frame then the process repeats at Step 1 where the error buffer in the compression processor is read. That error is accumulated in the error buffer at Step 2 and that accumulated error is compared to threshold at Step 3. Note that this looping will continue from Steps 1, 2, 3, and 4 until at Step 3 the accumulated error exceeds the threshold at which point it is not longer necessary to check any more microblocks for that frame because the error became so high that the host computer determined that a new I frame should be sent to restart the motion sequence. If, on the other hand, the accumulated error for all the microblocks of an entire frame never exceeds the threshold, then after Step 4, the process will go to Step 5 and the standard MPEG compression process will continue without changes, i.e., the next B or P frame will be grabbed and compressed.
 The ISMP still image compression methodology of the present invention can be used to greatly enhance automatic target recognition systems because the invention emphasizes and accurately represents the natural features such as “sculpture” of the object that makes human cognition of the target easier and more accurate. Furthermore, the polynomials used to represent the sculpture of the object are stable for small variations of projection direction or changes in movement, rotation, and scale of an object. This, too, enhances automatic target recognition.
 Human vision defines objects principally by their contours. The human visual system registers a greater difference in brightness between adjacent visual images that are registered, faithfully recording the actual physical difference in light intensity. Researchers have shown that individual neurons in the visual cortex are highly sensitive to contours, responding best not to circular spots but rather to light or dark bars or edges. At it turns out, the fact that ISMP compression extracts exactly these edges and emphasizes the “sculpture” characteristic of the object makes it especially advantageous for use in ATR. By preserving object edges in the compressed information, the human visual system can extract important objects from a background even if the object has bulk colors very close to the colors of other objects in the background. This feature is extremely important for registration of military targets.
 In virtually all ATR applications, the structures to be identified have sculpture. Consequently, the sculpture portion of the image can be extracted using the inventive methodology to achieve compression ratios of at least 4:1. Unlike prior art methods based on linear methods and Fourier transforms like JPEG and wavelet, which destroy the very information which is essential for human cognitionsoft edges, the present invention preserves those soft edges that exist in sculptures in virtually all structures to be identified. In contrast, the “texture” of an object is far less critical to human cognition. The present invention takes advantage of the distinction between sculpture and soft edges and texture by separating the sculpture characteristics of the object from the texture characteristics and utilizing only the sculpture information for ATR. An additional benefit of this methodology is that the sculpture information may be transmitted using relatively little bandwidth because it can be fully represented by polynomials whereas texture information requires greater bandwidth.
 A preferred method of ATR involves separating the texture and sculpture portions of the image using the ISMP compression method, using standard soft ATR on the sculpture portion, and then using standard hard ATR methods on the entire image (both texture and sculpture). Another preferred method for ATR in accordance with the presence invention to split the texture and sculpture portions of the image using a portion of the ISMP compression method, using state of the art soft ATR methods on the sculpture part, and then using state of the art hard ATR methods on the sculpture part. This greatly reduces the number of bits that need to be transmitted because the texture information is dropped altogether. Quality, however, remains high because the sculpture portion of the image was derived using ISMP which retains all necessary soft edge information which is critical to human cognition. Such soft edge information would be eliminated or lost, in any event, if standard Fourier transform type compression methods are used.
 There are numerous applications of ATR using datery obtained from the ISMP method. The datery can be used for autonomous object target detection, tracking, zooming, image enhancement, and almost realtime early stage recognition purposes. The present invention provides the capability for smart networkbased cooperative scene processing such as in remote intelligent consolidated operators (“RICO”) where information from remote camera networks must be transmitted over a smart local area network (LAN) which interconnects a number of camera platforms for cooperative wide area surveillance and monitoring. For example, a camera platform (with the inventive ISMP method embedded therein) can extract features of the objects seen such as critical soft edge information. It can transmit those images over a smart LAN to adjacent camera platforms. This process may provide cooperative scene information transmission outside the coverage of the original or any single camera platform. Through this process, observers of a scene can perceive the “big picture”.
 The images must then be transmitted from the remote camera network to a central station which may provide editing of film by computer to create the big picture. The invention will benefit such a system in two ways. First, because the ISMP compression method emphasizes the sculpture characteristics of the object, it enhances the ability to recognize the object imaged. Second, because the “sculpture” characteristics of the object are emphasized and represented using discrete numbers or coefficients from polynomials, the data sent is highly compressed which increases bandwidth significantly.
 Another application is as an autonomous movie director where standard ATR is used and that information is compressed using the present invention for sending those images from the cameras to the central station. Because of the large volume of information that can be generated in such a system, the images must be compressed sufficiently so that they do not overwhelm the host computer. This is a real problem that is solved by the hypercompression of the present invention. These benefits apply to a wide range of systems including battlefield imaging systems and antiterrorist recognition applications as well as full mapping capabilities.
 Another significant advantage of the present invention is the ability to provide sufficiently high compression ratios for providing TVclass transmission through traditional air communication channels which are 64 kbps or less. In fact, the invention can provide such a significant compression ratio improvement of more than an order of magnitude that, generally speaking, “video through audio” is made possible. In other words, the present invention makes it possible for battlefield commanders and others to receive image information as opposed to raw data. And because the image information they receive is sent in the form of discrete numbers or coefficients of polynomials that relate to isomorphic singular manifolds in the object, the data are highly compressed. And although highly compressed, the data preserve full information about the objects 3D boundaries or soft edges.
 An example of a realtime remote engagement (RTRE) air scenario made possible because of the present invention includes providing an aviator who is approaching his target (at sea, on the ground, or in the air) with a short TVrelay from an overflying military communications aircraft or satellites that upgrades the present target location at the last minute. This can prevent an aviator from losing track of a highly mobile target. This is made possible because the data are highly compressed and can be sent over low bandwidth air channels of 64 kbps or less and because the information that is sent preserves edge information which makes it possible for the pilot to easily recognize his target.
 Because the typical air communication channels are of low bandwidth, the ability to use all that bandwidth is critical. The present invention's ability to dynamically allocate bandwidth on demand permits the use of small fractions of standard 64 kbps bandwidths for bursty compressed video/graphic image transmission. A typical air communications channel must accommodate signals of different types such as imagery, audio, sensory data, computer data and synch signals etc. The higher level protocols of the network will prioritize these different signals. Conservatively speaking, imagery is one of the lowest priority because in most cases operations can continue without it. Therefore, imagery information typically is relegated to using only the bandwidth that is available and that available bandwidth changes with time. It is extremely useful to use the ISMP method of the present invention which can be implemented with a tunable compression ratio. This is distinct from software which changes compression ratio based on the type of the object. Furthermore, intelligence systems such as genetic algorithms or fuzzy logic and neural networks can provide intelligent control of the available bandwidth and permit imagery data to be sent where otherwise it was not possible to do so.
 The severe constraints placed on the tradeoff between the compression ratio and the PSNR by standard air channels of 64 kbps or less are highlighted by the following example. To compress data into the required data rate of 64 kbps from a fully developed synthetic aperture radar (SAR), for instance, uncompressed bandwidth of 13 Mbps, a 203:1 still image compression rate is needed. (512)^{2 }number of pixels, 10bit grey level, and 5 Hz bursty frame rate, yields (512×10×5=13 Mbps). The situation is made even more severe for VGA fullmotion video (221 Mbps) which requires a 3452:1 motion video compression rate. The ability of the motion estimator of the present invention which inserts I frames only where the content of the video requires it can provide ten times better compression ratios than prior art systems, namely, up to 4000:1. Thus, signals from an SAR uncompressed bandwidth of 13 Mbps may in fact, for the first time be sent through 64 kbps channels. This is made possible by the nonintuitive use of Arnold's Theorem according to which local isomorphism (i.e., 1:1 direct and inverse relation) exists between the 3D object boundary and its 2D image. As a result, the most critical part of the object—its 3D boundary may be described only by a 1D contour and 3 or 4 natural digits that characterize a simple catastrophic polynomial. This creates tremendous lossless compression of object boundaries and still preserves high image quality. Experimental results show that the ISMP methodology of the present invention in contrast to state of the art compression methods can achieve a compression ratio of 160:1 at PSNR=38 Db with almost invisible artifacts whereas the prior art offers only CR=60:1 at a lower PSNR value of 26 Db. The difference in the image is significant.
 Where far less than 64 kbps bandwidth channels are available, the hypercompression made possible by the present invention can permit continuity of images by “cartooning” which allows transmission of reduced realtime videoconferencing even on an 8 kbps communications channel.
 Referring now to FIG. 19, FIG. 19 shows five categories of data reduction as a fraction of the original, five reduced data rates, and five different outcomes of those data rates. Category A represents 100% of the original which is a data rate of 64 kbps. At this data rate, the original video may be sent. Category B represents a data reduction as a fraction of the original 75% which is a 48 kbps data rate. The result of this transmission is that tiny details of the face or other structure are still recognizable and edges remain unchanged. Category C represents a data reduction of 50% as a fraction of the original which is a data rate of 32 kbps. The result of this transmission is that edges are hardened and there are smooth transitions for face details. Category D represents a 25% data reduction as a fraction of the original which is a reduced data rate of 16 kpbs. The result of this transmission is a heavily reduced texture and hard edges but it is still possible to recognize a human face. Category E represents a 10% reduction in data as a fraction of the original which is a reduced data rate of 12.8 kpbs. The result of this transmission is hard edges and “cartoon” type faces. While cartooning certainly does not provide optimum viewability, it may be more than adequate, for example, for soft ATR purposes where a tank need only be distinguished from a plane or other categories of objects and the type of model of each is not required to be determined. Additionally, it was not possible to send even cartoon type images over low bandwidth communication channels using prior art methods and therefore the ability to send a cartoon type image over that communication channel where no image was possible before is a great advance. Thus, depending upon the quality of transmission required and the application, the compression techniques of the present invention may be utilized to achieve a broad array of results heretofore unobtainable using prior art compression methods.
 Various modes of carrying out the invention are contemplated as being within the scope of the following claims particularly pointing out and distinctly claiming the subject matter which is regarded as the invention.
Claims (31)
1. A method of compressing an image said method comprising the steps of:
segmenting said image;
creating a modeled surface for each segment;
connecting the segments to create an entire modeled image;
generating a texture image; and
combining said texture image and entire modeled image.
2. The method of claim 1 wherein the step of creating a modeled surface for each segment further comprises the step of choosing a canonical polynomial to represent isomorphic singularities in said image.
3. The method of claim 1 further comprising the steps of applying lossy compression to said texture image; and
applying lossless compression to the combined said texture image and said entire modeled image.
4. A method of compressing an image, I_{O }said method comprising the steps of:
dividing said image I_{O }into segments, each segment having a plurality of pixels;
calculating the dynamic range of the pixels in each segment;
selecting a best match canonical polynomial for each of said segments;
finding substitutes for variables in said canonical polynomial to calculate a modeled surface equation F for each of said segments;
creating a modeled surface I_{M }for each of said segments by substituting the coordinates of each pixel into the modeled surface equation F;
storing the coefficients for the modeled surface equation F for each of said segments;
finding connections between adjacent ones of said segments;
creating an entire modeled image, I_{M}, from each of said segments;
finding the difference between the image I_{O }and the entire modeled image I_{M }to create a texture image I_{d};
applying standard lossy compression to said texture image I_{d};
storing the texture image I_{d};
combining the entire modeled image I_{M }and the texture image I_{d }to create a combination image, and
applying lossless compression to said combination image.
5. A method of compressing a still image said method comprising the steps of:
identifying at least one catastrophe in said image;
representing said catastrophe with a canonical polynomial;
transforming said canonical polynomial into datery.
6. A method of compressing a still image, said method comprising the step of:
representing a region of abrupt changes in pixel intensity in said still image with a canonical polynomial.
7. A method of compressing an image, said method comprising the steps of:
segmenting the image into blocks of pixels;
creating a canonical polynomial surface for at least one catastrophe in at least one of said blocks of pixels;
sending the coefficients of said at least one canonical polynomial as compressed data.
8. A method of compressing an image, said method comprising the steps of:
identifying at least one isomorphic singularity in said image by applying photometric projection to said image;
characterizing said at least one isomorphic singularity with at least one polynomial.
9. The method as defmed in claim 8 , further comprising the step of:
creating a modeled surface of said image with said at least one polynomial, said modeled surface being isomorphically related to said image.
10. The method of claim 8 further comprising the step of:
transmitting the coefficients of said at least one polynomial as compressed data.
11. A method of compressing an image having manifolds, said method comprising the steps of: modeling the image as a photometric projection of at least one manifold in said image;
mapping said at least one manifold in coordinates (x,y,B) where the coordinate B is luminance at each point (x,y);
characterizing the mapping with a polynomial, said polynomial having coefficients;
sending the coefficients of the polynomial as compressed data.
12. The method of claim 11 , wherein the coordinate B does not introduce new singularities.
13. A method of compressing an image, said method comprising the step of describing the shape of object boundaries in the image in polynomial form.
14. A method of compressing an image, said method comprising the steps of:
segmenting the image into segments;
creating a modeled surface for each segment, said modeled surface for each segment being isomorphic with respect to each segment;
connecting adjacent segments to create an entire modeled image, said entire modeled image being isomorphic with respect to said image.
15. The method as defmed in claim 14 , further comprising the steps of:
calculating the peak signal to noise ratio over the entire modeled image;
calculating the difference between said image and said entire modeled image to retrieve texture information of said image.
16. A method of compressing video, said method comprising the steps of;
determining the error between a current frame and a predicted frame;
inserting an I frame as the next subsequent frame after the current frame if said error exceeds a predetermined threshold.
17. A method of compressing video, said method comprising the steps of:
taking a frame F_{0 }of the video;
segmenting the frame F_{0 }into search blocks;
predicting a subsequent frame;
determining the error between the frame F_{0 }and said predicted frame;
comparing said error to a threshold;
inserting an I frame as the next subsequent frame if said error exceeds said threshold.
18. The invention as defmed in claim 17 , further comprising the steps of segmenting the frame F_{0 }into microblocks within said search blocks and normalizing the error by dividing by the number of microblocks.
19. The method as defmed in claim 17 , wherein said threshold is based upon video content.
20. The method as defmed in claim 17 , wherein said I frame is inserted in place of said frame F_{0 }if said error exceeds said threshold.
21. The method as defmed in claim 17 , further comprising the step of dynamically changing the compression ratio on a frame by frame basis based upon said error.
22. The invention as defmed in claim 21 , wherein one of a genetic algorithm, neural network, and fuzzy logic are used to determine the necessary change in compression ratio.
23. A method of compressed video transmission, the method the steps of:
taking a first frame F_{O};
segmenting said frame F_{O }into blocks;
compressing each said block by representing it with a canonical polynomial having original coefficients;
predicting a frame P subsequent to said frame F_{O};
determining the error between each block in said frame F_{O }and said predicted frame P;
accumulating the error determined between each block in said frame F_{O }and said predicted frame P;
comparing said accumulated error to a threshold;
if said accumulated error exceeds said threshold, inserting an I frame as the next subsequent frame to said frame F_{O}.
24. A method as defmed in claim 23 wherein said I frame was previously compressed by representing it with canonical polynomials.
25. A method of compressed video transmission, the method comprising the steps of:
taking a first frame F_{O};
segmenting said frame F_{O }into blocks;
compressing each said block by representing it with a canonical polynomial having original coefficients;
predicting a frame P subsequent to said frame F_{0};
comparing each block in said frame F_{0 }with said predicted frame P to determine if there is a match;
if a match is found in said frame P for a block in said frame F_{0}, sending said coefficients of said polynomial for that block to a decoder;
if a match is not found for a block in said frame F_{0}, generating new coefficients of said polynomial representing said predicted frame P and sending said new coefficients to said decoder;
reconstructing said frame F_{0 }in said decoder from said original coefficients and said new coefficients sent to said decoder;
calculating an error between said frame F_{0 }and said predicted frame P based on said comparing each block in said frame F_{O }with said predicted frame P;
comparing said error to a threshold;
if said error does not exceed said threshold, sending a B or P frame as the next subsequent frame to said frame F_{O};
if said error exceeds said threshold, inserting an I frame as the next subsequent frame to said frame F_{O}.
26. The method as defmed in claim 25 , further comprising the steps of replacing said original coefficients, and said new coefficients with an I frame.
27. A method of compressing a video image, said method comprising the steps of:
taking a frame F_{O};
segmenting F_{O }into blocks and defining motion vectors for said blocks to predict a subsequent frame P having corresponding blocks;
defining errors between said blocks in said frame F_{O }and said corresponding blocks in said frame P;
accumulating said errors; and
based on said accumulated error, sending an I frame as the next subsequent frame to said frame F_{O}.
28. A method of automatically recognizing a target in an image, said method comprising the steps of:
segmenting said image into segments;
creating a modeled surface for each segment;
connecting the segments to create an entire modeled image; and
comparing said entire modeled image to a library of known images to determine if there is a match with a known image in said library of images.
29. The method of claim 28 wherein the step of connecting said segments to create said entire modeled image produces a target image having the sculpture characteristics of the image.
30. A method of automatic target recognition, said method comprising the steps of:
taking an image having texture and sculpture characteristics;
processing the image so as to separate said texture characteristics from said sculpture characteristics;
comparing said sculpture characteristics to a library of known images.
31. A method of analyzing an image for automatic target recognition, said method comprising the steps of:
separating the image into texture components and sculpture components;
applying soft ATR to said sculpture components to create a soft ATR sculpture component;
combining said texture components and said soft ATR sculpture components to form a combined image;
applying hard ATR to said combined image.
Priority Applications (2)
Application Number  Priority Date  Filing Date  Title 

US08/901,832 US6167155A (en)  19970728  19970728  Method of isomorphic singular manifold projection and still/video imagery compression 
US09/745,354 US20020176624A1 (en)  19970728  20001221  Method of isomorphic singular manifold projection still/video imagery compression 
Applications Claiming Priority (1)
Application Number  Priority Date  Filing Date  Title 

US09/745,354 US20020176624A1 (en)  19970728  20001221  Method of isomorphic singular manifold projection still/video imagery compression 
Related Parent Applications (1)
Application Number  Title  Priority Date  Filing Date  

US08/901,832 Division US6167155A (en)  19970728  19970728  Method of isomorphic singular manifold projection and still/video imagery compression 
Publications (1)
Publication Number  Publication Date 

US20020176624A1 true US20020176624A1 (en)  20021128 
Family
ID=25414887
Family Applications (4)
Application Number  Title  Priority Date  Filing Date 

US08/901,832 Expired  Lifetime US6167155A (en)  19970728  19970728  Method of isomorphic singular manifold projection and still/video imagery compression 
US09/745,363 Active 20230924 US7747085B2 (en)  19970728  20001221  Method of isomorphic singular manifold projection still/video imagery compression 
US09/745,392 Expired  Lifetime US6487312B2 (en)  19970728  20001221  Method of isomorphic singular manifold projection still/video imagery compression 
US09/745,354 Abandoned US20020176624A1 (en)  19970728  20001221  Method of isomorphic singular manifold projection still/video imagery compression 
Family Applications Before (3)
Application Number  Title  Priority Date  Filing Date 

US08/901,832 Expired  Lifetime US6167155A (en)  19970728  19970728  Method of isomorphic singular manifold projection and still/video imagery compression 
US09/745,363 Active 20230924 US7747085B2 (en)  19970728  20001221  Method of isomorphic singular manifold projection still/video imagery compression 
US09/745,392 Expired  Lifetime US6487312B2 (en)  19970728  20001221  Method of isomorphic singular manifold projection still/video imagery compression 
Country Status (5)
Country  Link 

US (4)  US6167155A (en) 
EP (1)  EP1000504A2 (en) 
JP (1)  JP2001511627A (en) 
KR (1)  KR100451942B1 (en) 
WO (1)  WO1999005851A2 (en) 
Cited By (24)
Publication number  Priority date  Publication date  Assignee  Title 

US20020143556A1 (en) *  20010126  20021003  Kadatch Andrew V.  Quantization loop with heuristic approach 
US20030072470A1 (en) *  20011015  20030417  Lee Henry C.  Two dimensional autonomous isotropic detection technique 
US20030115050A1 (en) *  20011214  20030619  Microsoft Corporation  Quality and rate control strategy for digital audio 
US6615158B2 (en) *  20010625  20030902  National Instruments Corporation  System and method for analyzing a surface by mapping sample points onto the surface and sampling the surface at the mapped points 
US20040001634A1 (en) *  20020628  20040101  Microsoft Corporation  Text detection in continuous tone image segments 
US20040001638A1 (en) *  20020628  20040101  Microsoft Corporation  Rate allocation for mixed content video 
US20040017939A1 (en) *  20020723  20040129  Microsoft Corporation  Segmentation of digital video and images into continuous tone and palettized regions 
US20040189863A1 (en) *  19980910  20040930  Microsoft Corporation  Tracking semantic objects in vector image sequences 
US20040213475A1 (en) *  20030426  20041028  Grohs Randall E.  Compression of images using object maps 
US20050015259A1 (en) *  20030718  20050120  Microsoft Corporation  Constant bitrate media encoding techniques 
US20050015246A1 (en) *  20030718  20050120  Microsoft Corporation  Multipass variable bitrate media encoding 
FR2874768A1 (en) *  20040831  20060303  Ecole Polytechnique Etablissem  Signal compression method 
US20070208548A1 (en) *  20060303  20070906  Solido Design Automation Inc.  Modeling of systems using canonical form functions and symbolic regression 
US7710452B1 (en) *  20050316  20100504  Eric Lindberg  Remote video monitoring of nonurban outdoor sites 
US20100111162A1 (en) *  20081030  20100506  Vixs Systems, Inc.  Video transcoding system with drastic scene change detection and method for use therewith 
US7925774B2 (en)  20080530  20110412  Microsoft Corporation  Media streaming using an index file 
US20110138154A1 (en) *  20091208  20110609  International Business Machines Corporation  Optimization of a Computing Environment in which Data Management Operations are Performed 
US8189666B2 (en)  20090202  20120529  Microsoft Corporation  Local picture identifier and computation of colocated information 
US8254455B2 (en)  20070630  20120828  Microsoft Corporation  Computing collocated macroblock information for direct mode macroblocks 
US8265140B2 (en)  20080930  20120911  Microsoft Corporation  Finegrained clientside control of scalable media delivery 
US8325800B2 (en)  20080507  20121204  Microsoft Corporation  Encoding streaming media as a high bit rate layer, a low bit rate layer, and one or more intermediate bit rate layers 
US8379851B2 (en)  20080512  20130219  Microsoft Corporation  Optimized client side rate control and indexed file layout for streaming media 
WO2015170333A3 (en) *  20140507  20160519  Sankhya Technologies Private Limited  Synthesis of a manifold system from input models 
WO2017204886A1 (en) *  20160523  20171130  Massachusetts Institute Of Technology  System and method for providing realtime superresolution for compressed videos 
Families Citing this family (54)
Publication number  Priority date  Publication date  Assignee  Title 

US7469237B2 (en) *  19960502  20081223  Cooper David L  Method and apparatus for fractal computation 
US6167155A (en) *  19970728  20001226  Physical Optics Corporation  Method of isomorphic singular manifold projection and still/video imagery compression 
US6330025B1 (en) *  19990510  20011211  Nice Systems Ltd.  Digital video logging system 
US6633837B1 (en) *  19991014  20031014  Object Reservoir  Method and system for generating software code using a symbolic language translator 
JP4810052B2 (en)  20000615  20111109  オートモーティブ システムズ ラボラトリー インコーポレーテッド  Occupant sensor 
US6384745B1 (en) *  20001010  20020507  Morisawa & Co., Ltd.  System, method and computer program product for a highly compressed outline font engine 
US7248257B2 (en) *  20010214  20070724  Technion Research & Development Foundation Ltd.  Low bandwidth transmission of 3D graphical data 
US20020135695A1 (en) *  20010326  20020926  Edelson Steven D.  Video data reduction by selected frame elimination 
US7020672B2 (en) *  20010330  20060328  Koninklijke Philips Electronics, N.V.  Reduced complexity IDCT decoding with graceful degradation 
US6968073B1 (en)  20010424  20051122  Automotive Systems Laboratory, Inc.  Occupant detection system 
WO2002088690A1 (en) *  20010430  20021107  Prabhu Nagabhushana  Optimization on lie manifolds 
DE60103280T2 (en) *  20011011  20050519  Alcatel  Apparatus and method for transmitting a gigabit Ethernet signal through a highcapacity pointtopoint radio system 
US7248741B2 (en) *  20020109  20070724  Hiroshi Akimoto  Video sequences correlation and static analysis and scene changing forecasting in motion estimation 
US20040001548A1 (en) *  20020422  20040101  Mcintyre Kristen A.  Method and apparatus to facilitate variableloss image compression for images of objects in motion 
US20050131660A1 (en) *  20020906  20050616  Joseph Yadegar  Method for content driven image compression 
US20040146211A1 (en) *  20030129  20040729  Knapp Verna E.  Encoder and method for encoding 
US20040175056A1 (en) *  20030307  20040909  Chulhee Lee  Methods and systems for objective measurement of video quality 
US7406181B2 (en) *  20031003  20080729  Automotive Systems Laboratory, Inc.  Occupant detection system 
US8253750B1 (en) *  20040214  20120828  Nvidia Corporation  Digital media processor 
US7038618B2 (en) *  20040426  20060502  Budic Robert D  Method and apparatus for performing bistatic radar functions 
US7418145B2 (en) *  20040630  20080826  Matsushita Electric Industrial Co., Ltd.  Compression device and decompression device 
US20080263616A1 (en) *  20040701  20081023  Sami Sallinen  Method and Device for Transferring Predictive and NonPredictive Data Frames 
US7480417B2 (en) *  20041019  20090120  Microsoft Corp.  System and method for encoding mosaiced image data employing a reversible color transform 
US7983835B2 (en)  20041103  20110719  Lagassey Paul J  Modular intelligent transportation system 
US8184691B1 (en) *  20050801  20120522  Kevin Martin Henson  Managing telemetry bandwidth and security 
US7899114B2 (en)  20051121  20110301  Physical Optics Corporation  System and method for maximizing video RF wireless transmission performance 
US7925105B2 (en) *  20060314  20110412  Seiko Epson Corporation  Image transfer and motion picture clipping process using outline of image 
WO2007141721A1 (en) *  20060606  20071213  Nxp B.V.  Display device and method of providing illumination thereto 
US8711929B2 (en) *  20061101  20140429  Skyfire Labs, Inc.  Networkbased dynamic encoding 
US9247260B1 (en)  20061101  20160126  Opera Software Ireland Limited  Hybrid bitmapmode encoding 
US8224065B2 (en) *  20070109  20120717  Purdue Research Foundation  Reconstruction of shapes of objects from images 
WO2008092131A2 (en) *  20070125  20080731  Skyfire Labs, Inc.  Mobile device user interface for remote interaction 
US8798148B2 (en) *  20070615  20140805  Physical Optics Corporation  Apparatus and method employing preATRbased realtime compression and video frame segmentation 
KR101499545B1 (en) *  20071023  20150306  삼성전자주식회사  Digital image processing apparatus, method for controlling the same, recording medium storing program to implement the method, and method of compressing digital image 
US7821454B2 (en) *  20071120  20101026  Sirf Technology, Inc.  Systems and methods for detecting GPS measurement errors 
US8509552B1 (en) *  20090309  20130813  Marvell International Ltd.  Digital image processing error concealment method 
KR101019558B1 (en) *  20080923  20110308  샘솔정보기술(주)  Method for Compression in System for Editing Virtual Reality 
US8908984B2 (en)  20091005  20141209  I.C.V.T. Ltd.  Apparatus and methods for recompression of digital images 
WO2011042900A1 (en) *  20091005  20110414  I.C.V.T Ltd.  A method and system for processing an image 
US20110122224A1 (en) *  20091120  20110526  WangHe Lou  Adaptive compression of background image (acbi) based on segmentation of three dimentional objects 
US20110221901A1 (en) *  20100311  20110915  Gm Global Technology Operations, Inc.  Adaptive Scene Rendering and V2X Video/Image Sharing 
JP5507515B2 (en) *  20110914  20140528  富士フイルム株式会社  Nonreversible compression apparatus, operation control method thereof, and operation control program thereof 
US8417005B1 (en) *  20110915  20130409  Sunnybrook Health Sciences Centre  Method for automatic threedimensional segmentation of magnetic resonance images 
JP6191160B2 (en) *  20120712  20170906  ノーリツプレシジョン株式会社  Image processing program and image processing apparatus 
US9275302B1 (en) *  20120824  20160301  Amazon Technologies, Inc.  Object detection and identification 
WO2014074399A1 (en) *  20121107  20140515  Sony Corporation  Method and apparatus for tissue region identification 
US8942447B2 (en)  20121107  20150127  Sony Corporation  Method and apparatus for tissue region identification 
CA2905586A1 (en) *  20130315  20140918  Sean BURNETT  Remote trespassing detection and notification system and method 
CN106664439A (en) *  20140730  20170510  恩特里克丝有限公司  System for cloud streaming service, method for still imagebased cloud streaming service and apparatus therefor 
US9942031B2 (en)  20140903  20180410  Research Foundation Of The City University Of New York  System and method for privatekey fully homomorphic encryption and private search between rings 
JP6365253B2 (en) *  20141112  20180801  富士通株式会社  Video data processing device, video data processing program, and video data processing method 
US9693063B2 (en)  20150921  20170627  Sling Media Pvt Ltd.  Video analyzer 
US9749686B2 (en) *  20150921  20170829  Sling Media Pvt Ltd.  Video analyzer 
KR101760826B1 (en) *  20151119  20170725  한국건설기술연구원  Apparatus for Compressing the Reflectivity Data of Weather Radar and Method therefor 
Family Cites Families (40)
Publication number  Priority date  Publication date  Assignee  Title 

US4908874A (en) *  19800411  19900313  Ampex Corporation  System for spatially transforming images 
DE3177088D1 (en) *  19810410  19890921  Ampex  Controller for system for spatially transforming images 
US4463372A (en) *  19820324  19840731  Ampex Corporation  Spatial transformation system including key signal generator 
US4855934A (en) *  19861003  19890808  Evans & Sutherland Computer Corporation  System for texturing computer graphics images 
US5122873A (en) *  19871005  19920616  Intel Corporation  Method and apparatus for selectively encoding and decoding a digital motion video signal at multiple resolution levels 
US5018215A (en) *  19900323  19910521  Honeywell Inc.  Knowledge and model based adaptive signal processor 
US5384725A (en) *  19900518  19950124  Yale University  Method and apparatus for encoding and decoding using waveletpackets 
US5611038A (en) *  19910417  19970311  Shaw; Venson M.  Audio/video transceiver provided with a device for reconfiguration of incompatibly received or transmitted video and audio information 
JP3002019B2 (en) *  19910704  20000124  富士通株式会社  Image coding transmission system having a cell discard compensation function 
US5227878A (en) *  19911115  19930713  At&T Bell Laboratories  Adaptive coding and decoding of frames and fields of video 
DE69333896D1 (en) *  19920129  20051201  Mitsubishi Electric Corp  Apparatus and method for recording / reproducing video information 
US5379351A (en) *  19920219  19950103  Integrated Information Technology, Inc.  Video compression/decompression processing and processors 
JP2636622B2 (en) *  19920313  19970730  松下電器産業株式会社  Encoding apparatus and decoding apparatus of the coding method and decoding method and a video signal of the video signal 
US5862262A (en) *  19920330  19990119  The United States Of America As Represented By The Secretary Of The Navy  Method of encoding a digital image using adaptive partitioning in an iterated transformation system 
US5416856A (en) *  19920330  19950516  The United States Of America As Represented By The Secretary Of The Navy  Method of encoding a digital image using iterated image transformations to form an eventually contractive map 
US5353061A (en) *  19921008  19941004  International Business Machines Corporation  System and method for framedifferencing video compression/decompression using perceptuallyconstant information and image analysis 
US5329379A (en) *  19921022  19940712  International Business Machines Corporation  System and method of measuring fidelity of decompressed video signals and images 
IL104636A (en) *  19930207  19970610  Oli V R Corp Ltd  Apparatus and method for encoding and decoding digital signals 
JP2518503B2 (en) *  19930308  19960724  日本電気株式会社  Screen switching detection method 
JP2591441B2 (en) *  19930928  19970319  日本電気株式会社  Motion vector detection apparatus of moving image signal 
KR970003789B1 (en) *  19931109  19970321  조백제  Bit allocation method for controlling bitrate of video encoder 
JP3344601B2 (en) *  19931229  20021111  ソニー株式会社  Image data reproducing method and apparatus, still image detecting apparatus and method, scene change detection device and its method, and the frame synchronizer 
EP1139671B1 (en) *  19940325  20091216  Nec Corporation  Image reproducing device 
US5640468A (en) *  19940428  19970617  Hsu; ShinYi  Method for identifying objects and features in an image 
US5600375A (en) *  19940908  19970204  Intel Corporation  Rendering an inter verses intra video encoding decision based upon a vertical gradient measure of target video frames 
KR100235345B1 (en) *  19941229  19991215  전주범  Moving picture estimation apparatus and method in divided region 
US5644386A (en) *  19950111  19970701  Loral Vought Systems Corp.  Visual recognition system for LADAR sensors 
KR0154739B1 (en) *  19950419  19981116  김광호  Fractal image compression device and method 
DE19516743A1 (en) *  19950506  19961107  Philips Patentverwaltung  A process for reducing the image data by means of fractal image coding, as well as the encoder and decoder for implementing the method 
DE19516742A1 (en) *  19950506  19961107  Philips Patentverwaltung  Method for data reduction by means of fractal image coding 
US5835163A (en) *  19951221  19981110  Siemens Corporate Research, Inc.  Apparatus for detecting a cut in a video 
US5682204A (en) *  19951226  19971028  C Cube Microsystems, Inc.  Video encoder which uses intracoding when an activity level of a current macroblock is smaller than a threshold level 
US5724100A (en) *  19960226  19980303  David Sarnoff Research Center, Inc.  Method and apparatus for detecting scenecuts in a blockbased video coding system 
US5987459A (en) *  19960315  19991116  Regents Of The University Of Minnesota  Image and document management system for contentbased retrieval 
US5675382A (en) *  19960408  19971007  Connectix Corporation  Spatial compression and decompression for video 
EP0822516B1 (en) *  19960729  20051116  Matsushita Electric Industrial Co., Ltd.  Shape data compression method and apparatus 
US5793895A (en) *  19960828  19980811  International Business Machines Corporation  Intelligent error resilient video encoder 
US5751378A (en) *  19960927  19980512  General Instrument Corporation  Scene change detector for digital video 
JP3031613B2 (en) *  19961112  20000410  株式会社つくばソフト研究所  Color / grayscale image input output device and the input output method 
US6167155A (en) *  19970728  20001226  Physical Optics Corporation  Method of isomorphic singular manifold projection and still/video imagery compression 

1997
 19970728 US US08/901,832 patent/US6167155A/en not_active Expired  Lifetime

1998
 19980727 WO PCT/US1998/015962 patent/WO1999005851A2/en not_active Application Discontinuation
 19980727 KR KR20007000979A patent/KR100451942B1/en not_active IP Right Cessation
 19980727 EP EP19980938212 patent/EP1000504A2/en not_active Withdrawn
 19980727 JP JP2000504703A patent/JP2001511627A/en active Pending

2000
 20001221 US US09/745,363 patent/US7747085B2/en active Active
 20001221 US US09/745,392 patent/US6487312B2/en not_active Expired  Lifetime
 20001221 US US09/745,354 patent/US20020176624A1/en not_active Abandoned
Cited By (62)
Publication number  Priority date  Publication date  Assignee  Title 

US7088845B2 (en)  19980910  20060808  Microsoft Corporation  Region extraction in vector images 
US20040189863A1 (en) *  19980910  20040930  Microsoft Corporation  Tracking semantic objects in vector image sequences 
US7062445B2 (en)  20010126  20060613  Microsoft Corporation  Quantization loop with heuristic approach 
US20020143556A1 (en) *  20010126  20021003  Kadatch Andrew V.  Quantization loop with heuristic approach 
US6615158B2 (en) *  20010625  20030902  National Instruments Corporation  System and method for analyzing a surface by mapping sample points onto the surface and sampling the surface at the mapped points 
US7742622B2 (en)  20011015  20100622  Lockheed Martin Corporation  Two dimension autonomous isotropic detection technique 
US7298866B2 (en) *  20011015  20071120  Lockheed Martin Corporation  Two dimensional autonomous isotropic detection technique 
US20030072470A1 (en) *  20011015  20030417  Lee Henry C.  Two dimensional autonomous isotropic detection technique 
US20050143991A1 (en) *  20011214  20050630  Microsoft Corporation  Quality and rate control strategy for digital audio 
US7295971B2 (en)  20011214  20071113  Microsoft Corporation  Accounting for nonmonotonicity of quality as a function of quantization in quality and rate control for digital audio 
US7283952B2 (en)  20011214  20071016  Microsoft Corporation  Correcting model bias during quality and rate control for digital audio 
US20050143990A1 (en) *  20011214  20050630  Microsoft Corporation  Quality and rate control strategy for digital audio 
US7295973B2 (en)  20011214  20071113  Microsoft Corporation  Quality control quantization loop and bitrate control quantization loop for quality and rate control for digital audio 
US20050143993A1 (en) *  20011214  20050630  Microsoft Corporation  Quality and rate control strategy for digital audio 
US20050143992A1 (en) *  20011214  20050630  Microsoft Corporation  Quality and rate control strategy for digital audio 
US20050159946A1 (en) *  20011214  20050721  Microsoft Corporation  Quality and rate control strategy for digital audio 
US20050177367A1 (en) *  20011214  20050811  Microsoft Corporation  Quality and rate control strategy for digital audio 
US7277848B2 (en)  20011214  20071002  Microsoft Corporation  Measuring and using reliability of complexity estimates during quality and rate control for digital audio 
US7340394B2 (en)  20011214  20080304  Microsoft Corporation  Using quality and bit count parameters in quality and rate control for digital audio 
US7260525B2 (en)  20011214  20070821  Microsoft Corporation  Filtering of control parameters in quality and rate control for digital audio 
US20060053020A1 (en) *  20011214  20060309  Microsoft Corporation  Quality and rate control strategy for digital audio 
US20070061138A1 (en) *  20011214  20070315  Microsoft Corporation  Quality and rate control strategy for digital audio 
US7027982B2 (en)  20011214  20060411  Microsoft Corporation  Quality and rate control strategy for digital audio 
US20030115050A1 (en) *  20011214  20030619  Microsoft Corporation  Quality and rate control strategy for digital audio 
US7263482B2 (en)  20011214  20070828  Microsoft Corporation  Accounting for nonmonotonicity of quality as a function of quantization in quality and rate control for digital audio 
US7299175B2 (en)  20011214  20071120  Microsoft Corporation  Normalizing to compensate for block size variation when computing control parameter values for quality and rate control for digital audio 
US20060045368A1 (en) *  20020628  20060302  Microsoft Corporation  Rate allocation for mixed content video 
US20040001634A1 (en) *  20020628  20040101  Microsoft Corporation  Text detection in continuous tone image segments 
US20040001638A1 (en) *  20020628  20040101  Microsoft Corporation  Rate allocation for mixed content video 
US7200276B2 (en)  20020628  20070403  Microsoft Corporation  Rate allocation for mixed content video 
US6980695B2 (en) *  20020628  20051227  Microsoft Corporation  Rate allocation for mixed content video 
US7085420B2 (en)  20020628  20060801  Microsoft Corporation  Text detection in continuous tone image segments 
US7072512B2 (en)  20020723  20060704  Microsoft Corporation  Segmentation of digital video and images into continuous tone and palettized regions 
US20040017939A1 (en) *  20020723  20040129  Microsoft Corporation  Segmentation of digital video and images into continuous tone and palettized regions 
US20040213475A1 (en) *  20030426  20041028  Grohs Randall E.  Compression of images using object maps 
US7146058B2 (en) *  20030426  20061205  HewlettPackard Development Company, L.P.  Compression of images using object maps 
US20050015259A1 (en) *  20030718  20050120  Microsoft Corporation  Constant bitrate media encoding techniques 
US20050015246A1 (en) *  20030718  20050120  Microsoft Corporation  Multipass variable bitrate media encoding 
US7343291B2 (en)  20030718  20080311  Microsoft Corporation  Multipass variable bitrate media encoding 
US7383180B2 (en)  20030718  20080603  Microsoft Corporation  Constant bitrate media encoding techniques 
FR2874768A1 (en) *  20040831  20060303  Ecole Polytechnique Etablissem  Signal compression method 
WO2006027475A1 (en) *  20040831  20060316  Ecole Polytechnique  Signal compression method 
US7710452B1 (en) *  20050316  20100504  Eric Lindberg  Remote video monitoring of nonurban outdoor sites 
US20070208548A1 (en) *  20060303  20070906  Solido Design Automation Inc.  Modeling of systems using canonical form functions and symbolic regression 
US8332188B2 (en) *  20060303  20121211  Solido Design Automation Inc.  Modeling of systems using canonical form functions and symbolic regression 
US8254455B2 (en)  20070630  20120828  Microsoft Corporation  Computing collocated macroblock information for direct mode macroblocks 
US8325800B2 (en)  20080507  20121204  Microsoft Corporation  Encoding streaming media as a high bit rate layer, a low bit rate layer, and one or more intermediate bit rate layers 
US9571550B2 (en)  20080512  20170214  Microsoft Technology Licensing, Llc  Optimized client side rate control and indexed file layout for streaming media 
US8379851B2 (en)  20080512  20130219  Microsoft Corporation  Optimized client side rate control and indexed file layout for streaming media 
US8819754B2 (en)  20080530  20140826  Microsoft Corporation  Media streaming with enhanced seek operation 
US8370887B2 (en)  20080530  20130205  Microsoft Corporation  Media streaming with enhanced seek operation 
US7949775B2 (en)  20080530  20110524  Microsoft Corporation  Stream selection for enhanced media streaming 
US7925774B2 (en)  20080530  20110412  Microsoft Corporation  Media streaming using an index file 
US8265140B2 (en)  20080930  20120911  Microsoft Corporation  Finegrained clientside control of scalable media delivery 
US20100111162A1 (en) *  20081030  20100506  Vixs Systems, Inc.  Video transcoding system with drastic scene change detection and method for use therewith 
US8787447B2 (en) *  20081030  20140722  Vixs Systems, Inc  Video transcoding system with drastic scene change detection and method for use therewith 
US8189666B2 (en)  20090202  20120529  Microsoft Corporation  Local picture identifier and computation of colocated information 
US20110138154A1 (en) *  20091208  20110609  International Business Machines Corporation  Optimization of a Computing Environment in which Data Management Operations are Performed 
US8818964B2 (en)  20091208  20140826  International Business Machines Corporation  Optimization of a computing environment in which data management operations are performed 
US8554743B2 (en) *  20091208  20131008  International Business Machines Corporation  Optimization of a computing environment in which data management operations are performed 
WO2015170333A3 (en) *  20140507  20160519  Sankhya Technologies Private Limited  Synthesis of a manifold system from input models 
WO2017204886A1 (en) *  20160523  20171130  Massachusetts Institute Of Technology  System and method for providing realtime superresolution for compressed videos 
Also Published As
Publication number  Publication date 

US6487312B2 (en)  20021126 
WO1999005851A3 (en)  19990415 
US20010028743A1 (en)  20011011 
US6167155A (en)  20001226 
WO1999005851A2 (en)  19990204 
US7747085B2 (en)  20100629 
EP1000504A2 (en)  20000517 
KR100451942B1 (en)  20041008 
US20020015532A1 (en)  20020207 
JP2001511627A (en)  20010814 
Similar Documents
Publication  Publication Date  Title 

Bryt et al.  Compression of facial images using the KSVD algorithm  
Lewicki et al.  Probabilistic framework for the adaptation and comparison of image codes  
EP1548616B1 (en)  Features for retrieval and similarity matching of documents from the compressed domain  
US5892847A (en)  Method and apparatus for compressing images  
EP0873653B1 (en)  Featurebased video compression method  
US5699121A (en)  Method and apparatus for compression of low bit rate video signals  
Penna et al.  Progressive 3D coding of hyperspectral images based on JPEG 2000  
US7424157B2 (en)  Apparatus and method for processing image data  
Nakaya et al.  Motion compensation based on spatial transformations  
AU2005286786B2 (en)  Apparatus and method for processing video data  
US5598216A (en)  Method and apparatus for encoding/decoding a video signal  
Davoine et al.  Fractal image compression based on Delaunay triangulation and vector quantization  
Liu et al.  A twostep approach to hallucinating faces: global parametric model and local nonparametric model  
US6738424B1 (en)  Scene model generation from video for use in video processing  
EP0405754A2 (en)  A method of processing video image data for use in the storage or transmission of moving digital images  
EP1583357A2 (en)  Imaging system, image data stream creation apparatus, image generation apparatus, image data stream generation apparatus, and image data stream generation system  
US7548654B2 (en)  Headerbased scaling and cropping of images compressed using multiscale transforms  
US7006567B2 (en)  System and method for encoding threedimensional signals using a matching pursuit algorithm  
Lengyel  Compression of timedependent geometry  
US7457472B2 (en)  Apparatus and method for processing video data  
US20030076334A1 (en)  Texture replacement in video sequences and images  
Schuster et al.  RateDistortion based video compression: optimal video frame compression and object boundary encoding  
Li et al.  Progressive coding of 3D graphic models  
US8773536B1 (en)  Linear system based, qualitative independent motion detection from compressed MPEG surveillance video  
Peng et al.  Technologies for 3D mesh compression: A survey 
Legal Events
Date  Code  Title  Description 

STCB  Information on status: application discontinuation 
Free format text: ABANDONED  FAILURE TO RESPOND TO AN OFFICE ACTION 