TW200812393A - Method for data compression of computer-generated video holograms - Google Patents

Abstract

Method for data compression of a sequence of computer-generated video holograms, which are encoded as complex hologram values of a scene from image data with depth information, for the representation on a display device with a spatial light modulator (SLM), which generates a modulated wave field and thereby reconstructs the desired three-dimensional scene, where changes in limited sections of the image content to be encoded only cause changes in the hologram values in limited sections of the SLM, comprising the steps (S1) definition of a phase matrix with predetermined phase values, (S2) determining the visibility of the scene according to the current position of an observer and considering the depth information of the image content to be encoded, (S3) encoding the scene, where these visible object points are given the same defined phase value in each image of the sequence and irrespective of the actual depth information of the object point during a transformation into hologram values, such that in subsequent images unchanged object points imply unchanged hologram values, thus allowing efficient data compression of the sequence.

Description

200812393 TPPO/PUK206-011 Nine, invention description: [Technical field of invention] This case is a data compression computer to generate image holography operation method is used to process a sequence of computer-generated hologram images (CGVH, Computer-Generated Video Holograms). For standard holographic images, they are generally stored as photos or in other suitable interference formats. CGVH will use discrete holographic image data from a sequence of 3D scenes. There are sub-and storage, for example, electronic storage media such as computers in electronic components, I (4) processing H, connectors, or other hardware components. • The correction can produce interference and the existence of a complex two-dimensional wavefront pattern in the observer's eyewear, and the light wave recombines a three (four) image. The difference between the light wave and the front of the light wave will also be generated by a computer. (4) The image matrix of the omni-directional light adjustment module is coded so that the wave field will emit the self-material (4) 2 full image material value, which will be adjusted. Correction, and the pixels on the upper side of the screen will also disturb to reorganize the 3D scene. "The real-time reorganization of the dry hologram image required by the 会 domain will provide you with the body components and calculation methods:: The main difficulties and challenges of different important applications and images are I, 纟°, stomach, digital The omni-image is generally much larger than the traditional imagery. The image is too large, and it is easy to damage the storage media and similar. The amount of data, the 'channel components and busbar system' information 5 200812393 TPPO/PUK206 -011 Transmitting components. The transmission and processing of traditional image data has caused serious damage to resources. Therefore, various attempts and methods are trying to deal with and reduce the amount of data, for example, by data compression. The way to help high-performance storage media. This case proposes an efficient data compression method, which is an important tool to enhance the storage and transmission of data and ensure high efficiency. A batch of controllable data will be Increased acceptability and the form of CGVH system allocation. • [Prior Art] The calculated hologram image data will be encoded with a matrix of optical calibration modules. Together, the light-tuning parent-nuclear matrix is an electronically protected element that senses the entire amplitude and phase, in short, the complex amplitude or the simplified amplitude of the induced light. ^ In this case, the so-called "Light Modulator Matrix" and "Spatial Light Modulator" (SLM) are mainly used to control the intensity, by 10 switching, cutting, or adjusting A device for color and/or phase of a beam emitted by one or more independent light sources. A holographic image display device basically comprises a matrix of controllable pixels that can be tuned to penetrate the display panel The amplitude and / or phase of the recombination of all object points. For example, the SLM can be equipped with a raw light separation module (αομ, Ae〇ust〇〇〇ptic Discrete Modulator) or a continuous type of module. The specific implementation example of reorganizing the holographic image using the amplitude of the adjustment is a description of the advantages of the liquid helium display (LCD). This case is also related to other more capable devices. Included, used to correct enough continuous light to be a 6200812393 TPPO/PUK206-011 light wave front or a light wave contour. The so-called "pixel" refers to a controllable holographic image pixel on the SLM. : It represents a discrete value of a holographic image point and can be discretely controlled and addressed. Each pixel represents a holographic image point of a holographic image. In digital light processing (DLP) In the example, a pixel is a discrete, controllable microlens, or a small group of discrete, controllable microlenses. In the continuous SLM example, the pixel is a development area used to represent the holographic image. As for the color reconstruction example, pixels are basically subdivided into multiple sub-pixels to represent the three primary colors. In most hologram systems, the encoded hologram is a three-dimensional scene. The so-called "conversion" in this case can be interpreted broadly, including any mathematical or computer-aided approach that reproduces or approaches and is structured on Maxwellian waveform equations. But the most common is the Fourier transform, which can be easily programmed into the program and has a very accurate understanding of the optical system. 10 In the reorganization of the traditional two-dimensional image data, the light wavefront or scene of the three-dimensional object in the holographic image recombination is generated by the induction and overlap of continuous light waves. This means that the value of the holographic image can be measured and determined for each pixel, and only the wavefront is required to be reconstructed as each object point. Each single pixel will also provide recombination with the corresponding conditioned light to the rest of the scene. Conversely, each point of the scene is assigned its information to the rest of the hologram. If only a small area of a scene is changed when it is reorganized, the values of all pixels of the holographic image will be affected, and the new values will be overwritten to all 7 200812393 TPPO/PUK206-011 pixels for proper reconstruction. The entire object, even if the object has only a slight change. Changes in the object also produce the same effect as changes in the image points of the video frame, where the pixel values of the SLM are generated and are different from other pixels that remain unchanged in the scene. Therefore, we can clearly see that only insufficient compression ratio can produce a method with different image compression groups as described above. This also makes it difficult to compress such holographic images into different images. References In order to solve the above difficulties, many theories and papers have been proposed as countermeasures, for example, as data compression by KLT (Karhunen-Loeve transformation). In addition, Ishida's "addressing the bits in the conversion coding of holographic images" is also an efficient compression method for holographic images with a large amount of data. When compressing the surface, it is necessary to consider the reduction of the image data, that is, if necessary, the unnecessary part of the image must be discarded from the hologram image, for example by a band pass filter ( Bandpass filter) way. Unwanted parts will also be separated from the parts that must be left, which will effectively reduce the amount of data. The digital holographic image compression method proposed by T. Naughton for the reconstruction and recognition of three-dimensional objects mainly describes the three-dimensional object reorganization based on the phase-shifted holographic image. Loss of compression can be used. The standard uncorrupted compression technique will be used to store the real-time encoding of the separated data stream with 8 200812393 TPPO/PUK206-011 so that the true of the digital holographic image can be completed before the Burrows-Wheeler encoding algorithm With the virtual part. Lossable compression techniques are based on subsampling, quantization, and discrete Fourier transforms. Based on the above theory, the loss of quality error caused by the reorganization of holographic images cannot be recovered by frequently invoking the loss of data compression method. SUMMARY OF THE INVENTION In the theoretical claim put forward by the present invention, the main purpose is to solve the effective rate compression for a sequence of holographic image data in the instant recombination of the image hologram. The amount of data for each image that needs to be stored or processed must be substantially reduced. In addition, the results of holographic image data must also be used simply, or at least standardized data compression techniques, so that a resource-saving and economical compression ratio can be achieved. The theory proposed in this case is a sequence of image holographic data compression methods generated by a computer, wherein a sequence of image holograms generated by a computer is encoded in a format of complex holographic image values constructed on an image content, and Contains image data with deep cargo information. The sequence of lean material at this time is presented by a device that includes a spatial light calibration module, SLM, and a large number of pixels. A tuned wavefield is generated by SLM in sufficient continuous light, and the SLM is controlled by complex holographic image values, and the required real or virtual 3D scenes are recombined in space through induction. The Observer window will be generated in the recombination space and have the same trapezoidal interval as the 200812393 TPPO/PUK206-011 μ. The window will be located close to the viewer's eye and can track the exact observer position with a well-known position detection and tracking system. The theoretical framework of the case can see the area of the scene, and the trapezoidal reorganization of the scene from the SLM to the observer window. The trapezoidal area can be integrated into a pyramid shape because the observer window is much smaller than the SLM range. In the preferred embodiment of the case, the observer window is very small. If a step is performed in reverse, the changes in the subtle sections of the image content will be encoded and will only be caused by the SLM. In the finite section, the complex hologram image is changed. In other words, it is only in a few pixels. Step 1: Phase Matrix 6 In the first step of the traditional method, we first define a phase matrix 疋”) with a solid phase value. In a simple way, the random values of the phase will be evenly divided; however, we can also use the distribution of beans to make more improvements in image quality. If the solution of slm = degree has m pixels and n pixels, the phase matrix will suggest the same dimension. Step 2: Visibility In the second step of the method, the visibility of the scene is measurable. The y scene will be (4) by the object point s (i, which is seen in the (4) position and has the depth information. j) The structure of the structure. Through the early solution of the solid solution, the object in the trapezoidal region (10) will have a discrete structure, and it can measure which one, the object point of the image is an operation that can be seen from the observer window. The scene will be sliced into a number of virtual hierarchical segments, = 200812393 TPPO/PUK206-011 segments are parallel to the SLM. Such as * slm

It is better to have nXm (Π multiply by the same number of points, even if it is a big one, The same. & therefore can reach close to the observer: the two sections can be more clear than the section of the observer's scene to get more clear #女二, the simple concrete example proposed by the public will be able to check the rush point, Whether or not they are hidden behind a point close to the viewer's system. Other specific examples presented in this case are described in more detail in the following sections. If the matrix of the slice hierarchy is used to correspond to the SLM Matrix, the visibility check will reduce the amount from n*m visible object points to match the number of phase values in the first step. If we see that the number of object points is less than n*m The product, then all the remaining "," will be d and then zero. The scene will also exist in the structure of η * πι visible object points. The other ratios of the SLM matrix, the phase matrix and the visible object point matrix are It can be converted to each other. Step 3: The method known from the book is also structured in a more efficient compression theory for sequence data, that is, when the amplitude of the object point is not in the subsequent image of the sequence data. Change and maintain its phase value. The phase matrix defined in the second step can be used in each sequence of holographic images. In the process of converting to complex holographic image values, the visible object points will be Each sequence image is given the same defined phase value, 11 200812393 TPPO/PUK206-011 regardless of the exact depth information of the object point. If Lu-V is in the same subsequent image, then the result is Nagasaki It is given _ image value. If ~ object point is no longer in the household, weaving ~ Wang:: column 彳 == two him, change. No face will be treated in a traditional way. The sequence of grass image = image value is proposed. The compression method is also used, and the simple and fast jh grazing image compression method is additionally used. * His compression side == rule case h Kehart coding mode and two-phase coding mode It is further explained in the following chapters. Of course, other methods such as kinoform encoding can be used. Exhausted stone is assuming that a commercial and mass-produced SLM has a high-quality holographic image with a resolution of The premise is to effectively reduce the cost of storage media and data transmission components such as high price and demand. It is interesting to note that the specific implementation example proposed in this case is to provide high-volume mass production technology. > Of course, this case is a data compression computer to generate image holographic method, which is a data condensed for a sequence of holographic images generated by the computer, mainly to encode the image data with deep information into two pairs of scenes. Complex holographic image values so that it can be recombined and displayed in a display bullying with a = Spatial Light Modulator (SLM). This spatial light calibration module can generate a calibration correction. The three-dimensional scene required for reorganization, and the recombined scene will be solid for 2 12 200812393 TPPO/PUK206-011 image content for coding to become a full child m image value and there is a small & field material difference change in some parts of the SLM, which includes the following processing steps. (51) Define a phase matrix with a preset phase value Wi, j); (52) by means of a view The current position of the ritual comes to Lao Lu face Dan / you coffee ~ Xia Mu decides the visibility of the scene, and T considers the depth information of the shirt to be encoded in the structure of the visible object point j); θ (= encoding (4), where these visible objects are clicked) will be here

Columns (4) are set to the same defined phase value for each image ^ regardless of the true depth of the object point # whether the signal is converted to the holographic image value; j, the image will contain the omni image value that has not changed. An unaltered object point for effective data compression for the sequence information. As described in the data compression computer to generate image holography operation method: 糸, trapezoidal recombination space to produce a display device, wherein the slm will consume the base of the trapezoidal space, and the observer sees at least t It will cover the pupils who are obsessed with the eye and are used as the root of tracking the position of the actual observer. As described in the data compression computer to generate image holography, the area of the viewer window in $ will be small. Changes in the limited area of the valley within the encoded image will only result in a change in the complex holographic image value of the limited area of the SLM. As described in the data compression computer, the operation method of image holography is performed, and the objects in the eight objects are represented by the structure of the visible scene point S(i, j) of the two-axis discretization (along (4) such as 丨(9)). 13 200812393 TPPO/PUK206-011 As described in the data compression computer to generate image holography operation method, the object is represented by the structure of the visible scene point S (i, j) of the three-axis discretization in the recombination space . The data compression computer generates an image holography operation method, wherein the object is represented by the structure of the visible scene point S(i, j) which is discretized in the recombination space, and the recombination space at this time All of the segment layers used for discretization are parallel to the SLM and have the same number of rows and columns. The data compression computer produces an image holographic operation method as described, wherein the number of visible object points, the SLM dimension, the phase matrix, and the object discretization are all blame or the same. As described in the data compression computer to generate image holography operation method, the phase matrix dimension, SLM resolution, visible object point dimensions, and the matrix used as discretized segment layer, will be with SLM Parallel to each other and all of them are identical or identical. The data compression computer produces an image holographic operation method as described above, wherein if the number of visible object points is less than the specified dimension, the remaining points are designated as zero. The data compression computer generates an image holographic operation method as described above, wherein we only need to consider the visible object point S(i, j) when performing scene coding processing. The data compression computer produces an image hologram operation method in which the object point S(i, j) has its amplitude maintained to maintain a constant phase value in a subsequent sequence of images. 14 200812393 TPPO/PUK206-011 The data compression computer produces an image holographic operation method as described in which the phase matrix has an average value of the value of each row and/or the value of each column. As described in the data compression computer produces image holographic operations, where complex holographic image values are converted to Burckhardt components or a phase component. [Embodiment]

The theoretical techniques and compression devices proposed in this case will be further elaborated in the following paragraphs by way of specific examples. Examples of image holographic recombination devices that have been constructed in the past month have been made in the patent documents PCT/EP 20〇5 009604 and WO 2004/044659. The foregoing two methods and displays are constructed on the basis of not reconstructing the entire scene object, wherein the scene object can be seen by the observer, but will be projected into two small observer windows to cover The pupil of the observer's eye, and the wavefront emitted from the scene object and substantially present will be set at this position. The observer can observe the scene through the viewer window. The virtual observer window covers the pupil of the observer's eye and is able to track the exact position of the observer through the well-known position detection and tracking system. A virtual ladder recombination scene will be in the slm and observer windows of the holographic image display, where the SLM is the bottom of the trapezoidal area and the viewer window is the height of the disciple area. > If the observer window is very small, the resulting one will change to (4) as a pyramid. The observer will reproduce the wavefront of the scene through the virtual view and the window of the viewer, so the observer can see the recombination of the scene in the reorganization space. 200812393 TPPO/PUK206-011 is proposed in this case. In the first embodiment, the image-wide image generated by the computer is recombined by a holographic image display, and the holographic image display is already in the patent document PCT/EP 2005 009604, and the image generated by the computer The method and apparatus for holographic coding and recombination are described in the definition. The horizontal axis used to generate the fax effect can be obtained by means of a space work, such as using a lens mask and recombining the hologram image in the adjacent column of the SLM. This means that it is not reconstructed by the scene itself, but the wavefront that reproduces the wavefront itself or reproduces the scene will be produced within the small virtual observer window. Earth, the viewer window is so small, when we look back at the entire process, if you make changes to the changes made in the limited section of the image content, = will only lead to the complex hologram of the SML in the SLM The change and intersection of the image values will only affect a small number of pixels. Like Xin Xier Ming is based on the assumption that there is a certain SLM behavior of m 2 and 歹, η pixels. The SLM at this point is a slice LCD panel that can be calibrated to correct enough continuous (four) amplitude. The method of data compression proposed by this wood is to first define a ..., the mouth-phase (four) contains the phase matrix of the bit value in the first step. And the phase values are evenly distributed across the rows and columns. (4) With the number of columns (8), in the second step, the object points of the object of the scene are measurable. It is a radiochemical and can be measured and braked, and the sorrow is judged by the position of the observer. The so-called observer is the observation direction of the tiny observer window, and the object is visible. Consider the depth of the scene point 16 200812393 TPPO/PUK206-011 汛. In a simple way, the object will be cut into a number of fragment levels that are parallel to the SLM. In the trapezoidal region, each level of the object is scanned by the matrix of the fixed m彳η column. The m 4亍η column matrix is actually taken from the SLM and the resolution.纟 When considering the observer's eye position, no matter whether a new point is hidden behind the previous point, each point on the visible level must be checked. Because each pen data record has the same number of points as the matrix, the area to which the individual matrix is closed will change from the observer window. ▼See the inspection is mainly to find or select the servant mjn visible points, that is, for the visible object = f the case (4) the specific implementation of the checklist, the section level ί: special distance, but 疋 each other The distance between them may be connected by the deep sound of the scene: the news. With a tangible observation: the segments will be sent in each of the (four) segments and the calculated points 2 will also be judged by checking whether they are hidden near the view of the housekeeper. Cheng will be from the virtual nearest to the inspector =: These calculation points will be defined as the inner layer of the object and the inner level of the section, : in the eyes of the center of the burnt 2 window, a certain attraction is closer to the level The i-th row is the broadest: the brother column 'and hides in some if it is, then it will be removed from the current (four) layer: owed, then the processing level. In the third step, that is, the code is assigned to %T in each sequence image, "see the object point will be the same phase value of the beer, and 17 200812393 TPPO/PUK206-011 Whether the tribute is in the transition to complex holographic image values. If an object point is maintained in the same subsequent image, it will automatically assign the same complex hologram image value L to each sequence image so we can give a differential image compression so that a sequence of data Can be effectively reduced. In another specific embodiment proposed in the present case, if the spatial multiplexing is performed according to the above general, the observer window will be generated by means of time multiplexing, such as the patent document W〇2004/044659 "Recombined image. The hologram = image hologram and device" content. The two tiny bits of the image = perspective will be cast using the sequence switching between the light source and the SLM content, corresponding to the viewer window. The traditional data compression method can supply two image images at time 2. The amount of data stored will be further reduced by, for example, ignoring the differences in the perspective of the background objects, and the backgrounds in the two are individually independent. ^ Again... A holographic image value _ generated in the same way as the two specific embodiments described above will be encoded using Berkhardt encoding or two-phase encoding. If Berkhardt encoding is used, a complex value will be represented by three discrete, grayscale values, where the grayscale value ranges from 〇^255 and 255 represents the largest possible component. value. Based on a definition • The maximum possible signal amplitude, the true signal amplitude is normalized to the maximum value, and the available 255 steps are used as a reference. We — will pick the maximum value and initialize it at the beginning of the data compression. Changing the maximum value at the beginning of processing other images will lead to a large change in the value of the holographic image in all pixels, and also reduce the possible sequence data compression ratio. 〆 18 200812393 Maximum value of image sequence TPP〇/PUK2°6'011

The corresponding amount of Distribution) is not evenly distributed (Gaussian is the solution to cut off the 5 W early, and replaces the peak Gaussian distribution of the maximum value of the (1) π with the maximum value and Provide statistics to tolerate the corresponding choices - a higher maximum A value will be self-growth. - Generally speaking, when selecting the lower one, the larger one will only approximate the maximum value: this Take a large value of the day, a large number of sequence values will not be able to use the storage shirt corresponding to the wheel from the matrix, when the storage phase is like 20%, visible object points (including such as "calculation; piece: ====_ shadow data The partial sequence of the input data during compression is used as the method of Shi Shizeng 2 file de ι〇2_〇63 (4) mainly to describe a method for the image full image generated by =3⁄4. By this method, and =, Objects of complex amplitude values of the quasi-view will be assigned to parallel virtual segments: cuckoo arrays so that each segment layer will be defined as individual independent = material records with discrete amplitudes in the matrix points Value, and ^ as the holographic image display (four) The image code ^ is calculated from the image data record. This document also describes a single processing device that is used to generate the aforementioned method. The individual steps are 19 200812393 TPPO/PUK206-011 will be by this single processing device The corresponding components are generated one by one. The detailed steps are as follows: _ (S1) The process of converting the object data record of each segment layer into a two-dimensional distribution of the wave field in at least one virtual window, wherein the virtual window must exist in a reference Layers as close as possible to the observer's eyes; _(S2) aggregates the calculated distribution results for all segment layers in all reference records; - (S3) the reference records are converted to a position parallel to the reference layer and maintaining a finite distance The holographic image layer process, which is an SLM layer, produces a holographic image record that is used to assemble all computer-generated hologram images (CGH). The single processing device used for encoding is used in Producing a simple implementation of a traditional method. Conversely, a single processing device can also be modified by individual steps. To more extended applications. Due to the use of the segment layer, some corrections provide the opportunity to organize the scene into visible object points, as in the first specific implementation example mentioned above. The encoding will utilize the phase matrix The phase value is assigned to the visible object point extension, just like the specific implementation examples proposed in the foregoing. The holographic image generated by the computer will be divided into two-phase or three-phase components and compressed efficiently, for example in A differential image processing towel, which can also be generated by using a corresponding arithmetic component. The prior art is implemented by a person skilled in the art, and its unprecedented practice is also patented, and the embodiment is still It is not enough to cover the application of this article. However, the scope of the above-mentioned patents of the slaves of the woods, therefore, the application for 20 200812393 TPPO/PUK206-011 patent scope is attached.

21 200812393 TPPO/PUK206-011

[Simple description of the diagram] None. Simple description of component symbols: None 0 22

Claims (1)

200812393 TPPO/PUK206-011 Patent Application Scope 1 · A data compression computer produces image holography operation method, which is used for data compression of a sequence of holographic images generated by a computer, mainly encoding and processing image data with depth information. Become a complex holographic image value of a scene so that it can be recombined and displayed on a display device with a Spatial Light Modulator (SLM), and this spatial light calibration module can generate a tone After correcting the wave field and reorganizing the required three-dimensional scene, the reconstructed scene will have a small area error change in some parts of the SLM because the image content is encoded to become a holographic image value, which includes the following Processing steps: (51) defining a phase matrix having a preset phase value p(i, j); (52) determining the visibility of the scene by an observer's current position, and taking into account the depth information of the image content For encoding processing, the scene will be represented by the structure of the point S(i, j); the wide (53) will encode the scene, and these visible objects From 丨, 每 will be set to the _ defined phase value in each column of information in this column, and the true depth information of the object point is in the period of conversion to the holographic image value; '...what this will be imaged The unconformed object of the change of the wrapper f is effective for compressing the data for the sequence information. ",, 2. The application of the Wei-Electric operation method as described in item 1 of the patent application scope, which is used for one of the trapezoidal recombination spaces to produce the base of the trapezoidal space, and the observer sees the second to the ' The pupil of the SLM eye in the eye and the thoroughness of the position as the true _ position = Cover observation 23 200812393 TPPO / PUK206-011 = 2 The data of the second item is compressed to produce the image of the omnipotent window The area of the area will be small so that the change of the limited area of the valley in the image to be encoded will only result in the change of the complex holographic image value of the limited area. The data pressure described in J3 item _ brain image holography The method of operation, in which the object is presented by the structure of the visible view point S(i, j) of the three-axis discretization (discretisati (8). The information of the dedicated staff _ computer generated image holography in the recombination space three The visible data of the axis discretization ==== the data pressure described in the item _ produces the image holographic method #, /, the object of the towel is _ in the structure of the reorganization of the empty sound image sa D, the city time Reorganization (4) Discretized segment layer Read the SLM slanting line and butterfly _ the number of lines has and = for 7. As described in the application of the fourth paragraph: the hybrid compression operation method, the number of visible object points, the dimension of the SLM ^ and The object discretization is constant or the same. Phase matrix, 8 · As applied in the _ 丨 之 ( 四 四 压 魏 魏 魏 魏 魏 魏 魏 魏 魏 魏 魏 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中 其中The dimension of ^, and the matrix used as the discretized segment layer, will be parallel with the coffee phase 24 200812393 TPPO/PUK206-011, and they are all identical or identical. 9. For example, the fourth section of the patent application shed The operation method of _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Brain image full image rhrt method, which, when entering the scene coding process, we only need to consider the visible object 1 thousand points b | l, j). Cai 1 tear shape # material pressure _ vocational image hologram = Method, in which the object point s(i,]·) its amplitude will turn unchanged in a sequence of subsequent images Invariant phase value. mu1 _& fine_group imagery array whose value per line and/or value per column is flat mnr/r (4) compression fine to produce image hologram followed by Langekeha Special 25