CN1898680A - Systems and methods relating to afis recognition, extraction, and 3-d analysis strategies - Google Patents

Abstract

Provided is systems, methods for identification, extraction and three-dimensional analysis tactic of an AFIS system. Systems, methods, etc., that that assist a print examiner to thoroughly search and compare a print or substantial portion thereof against a known print database contained within an AFIS system. In certain embodiments, the prints can be definitively matched to a corresponding same print in the database. A result of a more thorough search and comparison can be a higher hit score and accuracy rate. In certain embodiments, the database comprises a candidate list of previously obtained prints to assist in the identification.

Description

System and method about identification, extraction and the three dimensional analysis strategy of automatic system of fingerprint recognition

The cross-reference of related application

The present invention requires the right of priority of following patented claim: the number of submitting on October 23rd, 2003 is 60/513,669 U.S. Provisional Patent Application; The number of submitting on November 6th, 2003 is 60/517,849 U.S. Provisional Patent Application; The number of submitting on November 7th, 2003 is 60/518,263 U.S. Provisional Patent Application; The number of submitting on February 27th, 2004 is 60/548,214 U.S. Provisional Patent Application; The number of submitting on April 14th, 2004 is 60/562,635 U.S. Provisional Patent Application; The number of submitting on May 19th, 2004 is 60/572,665 U.S. Provisional Patent Application; The number of submitting on June 23rd, 2004 is 60/582,414 U.S. Provisional Patent Application; And the number of submitting on August 23rd, 2004 is 60/604,092 U.S. Provisional Patent Application; It is all and for all instructions and openly be incorporated herein with for referencial use.

Background technology

Automatic system of fingerprint recognition (AFIS) system (http://onin.com/fp/afis/adis.html) can not search for traditionally with matching check person can be by all fingerprints of manual methods identification.In part because human potential marking overlooker can use additional detail, for example the Grade III details is mated fingerprint (as used herein, " marking " refers to animal and be generally the unique identification marking on the mankind, for example fingerprint, palmmprint, toe line, footmark etc.).Simultaneously, the physiological defect of human eye stops the overlooker to distinguish very meticulous grey measurement level rank, and it hinders the accuracy of the details mark (minutiae marker) in each of Grade I, II and III details in fingerprint and the identification of position.Rank 3 details are the meticulousst details in three ranks.Generally speaking, the wrinkle ridge line stream (ridge flow) of rank 1 finger mark note, the wrinkle ridge lines (ridge path) of rank 2 finger marks note, and rank 3 fingerprint ridge shapes, it comprises unique edge details, pore shape and position, wrinkle ridge shape and other details less than the wrinkle ridge width of beginning.Therefore, existing systems does not always allow AFIS system, operator and overlooker with characteristic explain be marked as details, and feature wherein can be by the nuance in the grey measurement level that manifests fingerprint image and distinguished exactly.

Not developing the AFIS system discerns, extracts, quantizes, searches for and mate Grade III feature in two or more fingerprint images.Also do not develop the AFIS system and can discern desired details mark with tag level III details.Do not having under the situation of this ability, can not exhaustive analysis the nuance of ash measurement level, so some feature may not be identified, extracts, quantizes, searches for and/or mates.In addition, AFIS III details can be present in different grey measurement level in the fingerprint image (or other magnitude for example form and aspect or saturation degree).If do not consider grey measurement level, some Grade III feature may not be identified, extracts, quantizes, searches for and/or mates so.(noticing that in fact AFIS III carries out or the Grade III analysis of the marking of execution manually with expression in this simple expression that is commonly used in automatic system).

Therefore, have unsatisfied needs for other AFIS system tool, this instrument participates between the feature that is present in different grey measurement levels and distinguishes, and this feature of mode mark can search for the AFIS system.Native system, method etc. provide these and other advantage.

Summary of the invention

In certain embodiments, native system and method are considered to be used for the AFIS3-D or the AFIS III+ of various embodiments at this, and help the overlooker for being included in the thorough search of the intrasystem known marking database of AFIS and comparing the marking or its essential part.Described essential part comprises that enough markings discern enough features and compare with search with the marking in given data storehouse, and in certain embodiments, comprise that enough markings come at least exploratoryly at least one coupling with the described marking and the known marking.In certain embodiments, the marking can be definitely with database in corresponding identical marking coupling.The result who searches for more completely and compare can have higher mark and the accuracy of finding.In certain embodiments, described database comprises the candidate list of the marking of previous acquisition, to help identification.

The more accurate comparison that is the potential marking and known marking image of this important benefits that provides (" marking " comprise any suitable body part for example lip, skin, fingerprint, keep the seal, the marking of toe seal etc.).For example, bomb debris (bomb fragment) can comprise only little local potential marking.AFIS 3-D and/or III+ system can more may cause the identification in potential marking source based on the search for the database of the marking/image more accurately.

This innovation is for example by using small gray scale in the image and/or other magnitude to search for this image/marking to produce more comprehensively and AFIS mark details mark (AFIS rank 2 or 3 marks) for example accurately in mode more accurately, and/or produce different paths (pathway) by using different magnitudes in the image repeatedly to search for this image/marking, described path is used for identification then, extract, quantize, search and mate comprehensive one group of Grade III feature, can help to discern and eliminate threat to home guard, and no matter described threat is abroad or domestic.

Therefore, in certain embodiments, comprise the analysis marking in this described method and related software and other system, it comprises: the marking images that at least 2 dimensions a) are provided; B) make described image stand magnitude and strengthen analysis, thereby at least one relative measurement at least one essential part of the described marking can be depicted in respect at least 2 another dimensions of tieing up, to provide magnitude to strengthen image, thereby than there not being described magnitude to strengthen 2 dimension images of analysis, other magnitude rank more can be discerned basically for human eye; C) show described enhancing image; And d) the described magnitude of manual examination (check) strengthens image, to place at least one details mark on the described marking.

Described put procedure comprise identification and place before on the described marking Unidentified at least one, two or more details marks, and can comprise and move at least one that before be placed on improperly on the described marking, two or more details marks.

Described method also can comprise the analysis marking, and it comprises: the marking image of at least 2 dimensions a) is provided, and it comprises the details mark of being determined by automatic details labeling algorithm, so that automated characterization point mark to be provided; B) make described image stand magnitude and strengthen analysis, thereby at least one the relative magnitude at least one essential part of the described marking is depicted in respect at least 2 another dimensions of tieing up, to provide magnitude to strengthen image, thereby than there not being described magnitude to strengthen 2 dimension images of analysis, in addition selected measurement or magnitude rank can be for human eye as seen; C) show described enhancing image; And d) the described magnitude of manual examination (check) strengthens image, to assess the correctness of described automatic details mark.Described method further can comprise at least one details mark of placing the described marking, and its described put procedure comprises to be removed incorrect automatic details mark, move incorrect automatic details mark or increase in the other details mark at least one.

On the other hand, comprise the analysis marking in this method, it comprises: the marking image of two dimension at least a) is provided; B) make described image stand magnitude and strengthen analysis, thereby at least one the relative magnitude at least one essential part of the described marking is depicted in respect to another two-dimentional at least dimension, to provide magnitude to strengthen image, thereby than there not being described magnitude to strengthen the two dimensional image of analyzing, other magnitude rank for human eye as seen; C) described magnitude is strengthened image and be divided into a plurality of intensity ranks; D) individually select at least one isolated intensity rank; E) determine the mark of at least one automatic system of fingerprint recognition from described isolated intensity rank.Described method further can comprise the described isolated rank of demonstration, and manually or automatically determines described at least one details mark.

Comprise in this application explained aspect these and other in the following detailed description and drawings, feature and embodiment.Unless specify in addition or become clear from context, all embodiment, aspect, feature etc. can mix in any desired way and mate, make up and change order.In addition, explained multiple reference at this, it discusses some system, device, method and out of Memory; All these are with reference to all and for all instructions with openly be incorporated herein doing reference, and no matter described reference appears at the position of this application.

Description of drawings

Fig. 1 is the screenshot capture (screenshot) according to the image of the marking of analyzing in this method.This image comprises 4 features of extracting automatically, and only 1 is correct.Other three features (on the right) are incorrect.

Fig. 2 depicts the potential marking identical with Fig. 1 that uses software to draw, and this software is pressed 3D surface rendering image strength characteristic.

Fig. 3-the 5th, a series of screenshot captures of the part of the image among Fig. 1 and Fig. 2, it demonstrates the dynamic figure of 3 dimension wrinkle ridge shapes, and wherein the z axle is represented grey measurement level.

Fig. 6 is normal screenshot capture that obtain and that keep the seal at the 2D of the level of intensity of another dimension shown in having.

Fig. 7 is to use the screenshot capture of the fingerprint that the embodiment of the 3D details tools of discussing at this draws.

Fig. 8 depicts through details and handles keep the seal identical with Fig. 6.

Fig. 9 depicts through rolling, tilting and keep the seal identical with Fig. 6 of shaking.

Figure 10 depicts by observed keep the seal identical with Fig. 6 of enhancement mode filtrator, this enhancement mode filtrator for example the contrast adjustment further to have increased vision clear.

Figure 11 depicts the keep the seal image identical with Fig. 6, and wherein the user has placed the details mark on the 3D surface, and with respect to existing 2D database, it can be quantized and search for.

Figure 12 a depicts the image of known marking impression.

Figure 12 b depicts wrinkle ridge lines that AFIS II system reads and based on the deviation of main wrinkle ridge lines.

Figure 12 c depicts the grades II details with direction and relation of AFIS II system identification and extraction.

Figure 13 a-c depicts the image identical with Figure 12 a-c with shown AFIS Grade III details.

Figure 14 depicts the AFIS III+ terrain analysis in the part of the grey measurement level of difference, identical with image among Figure 12 a-c marking.

The AFIS III+ that Figure 15 a-c depicts at the image identical with Figure 12 a-c of the grey measurement level of difference analyzes.

The AFIS III+ that Figure 16 a-c depicts with the grey measurement level different with the analysis of Figure 15 a-c, identical with Figure 12 a-c image analyzes.

Figure 17 is the table of expression from the progressive character of section (slice) combination of multi-level and many images.

Figure 18 depicts the image of the example of the recessed morphological feature that comprises in the marking image and protruding morphological feature.

Figure 19 depicts the image of the example that comprises the edge feature in the marking image.

Embodiment

AFIS 3-D method for visualizing and system

When the gray-scale value person of the being examined exhaustive analysis of gamut in the fingerprint image, the position of mark can be more accurate on the image that occurs as the part of AFIS systematic search.A series of marks are more accurate and/or place with bigger quantity than the analysis of the gray scale that may lack gamut or other magnitude, and it can increase the chance that AFIS finds.

Two problems of current AFIS analysis module (it does not allow the analyst to place the details mark based on the grey measurement level that is apparent on the 3-D face) are, all correlated characteristics that appear on the impression may the person of not being examined be discerned, or the accurate position of some features may be by correct not determining.Software and system overcome the human visual system in the weakness of distinguishing the gradation of image magnitude, and mode is for describing this value as the 3D surface, and it helps fingerprint detection person to distinguish the variation that the gradation of image magnitude is very trickle, and the AFIS mark is placed on correct place and position.

The AFIS system is generally with following dual mode identification details: the 1) layout and 2 of the fingerprint details mark of system's generation automatically) by analyst's arrangement details unique point mark.In addition, the overlooker can be placed on other mark on the marking image to replenish the mark that is produced automatically by the AFIS system, vice versa, and the overlooker is removable or delete incorrect mark on the marking image, to correct the incorrect mark that is produced automatically by the AFIS system.AFIS 3-D and AFIS III+ help human detection and details and details mark are placed in fingerprint image or other marking image, cause the AFIS Search Results that improves thus.

In Fig. 1, place the potential marking 30 of characteristic processing by the automatic extraction and the details of general indistinct finger mark workstation (Universal Latent Workstation, (ULW)).This software is used so that FBI IAFIS system is submitted in potential marking search by national law enforcement agency.Three minutia dot marks 32 on impression the right correctly do not reflect the structure of wrinkle ridge under the mark.It is incorrect.The unique point mark 33 on the left side is correct.By preferably dynamically (promptly with the 3-D face, wherein the 3-D face can roll, tilts and rock, be also referred to as and be vertical (pitch) partially, driftage (yaw) and/or rolling, and itself in addition can add cinema loop (cine loop), thereby the 3-D face is vertical partially with repetitive mode, driftage and/or roll) show not have this impression separately or to termination to advance in whole this zone along this regional wrinkle ridge lines.

In Fig. 2, draw out the same potential marking with software, this software is with 3D surface rendering image strength characteristic.Particularly, provide the digital picture of Fig. 1 to be used for drawing, image is subjected to intensity level enhancing analysis then, thereby at least one the relative intensity value at least one essential part of the marking is depicted in at least 2 another dimensions of tieing up.This provides intensity level to strengthen image, thereby ties up images than there not being intensity level to strengthen 2 of analysis, but other intensity level rank is for human eye more identification basically.But identification represents that in certain embodiments feature is sightless under not strengthening, and in other embodiments, strengthens down that feature is visible, but as having enhanced features, but feature more identification or identification more fast basically.By this characteristic of detailed inspection, many more accurate features are shown, and draw other AFIS unique point mark 34.For example, in Fig. 2,15 features are apparent on the dynamic 3D exterior view, and details mark 34 is painted on the image.

Fig. 3-the 5th, a series of screenshot captures of the part 36 of the image among Fig. 1 and Fig. 2, its dynamic (rock, tilt or roll) AFIS 3-D that 3 dimension wrinkle ridge shapes are shown analyzes, wherein the z axle is represented grey measurement level, this gives overlooker's more information and goes to manifest, and therefore having created chance places more details mark, and more accurate finishing.In some embodiment of this or others of the present invention, the part of image or its expectation also can be exaggerated (image amplification), dwindle etc.

When Fig. 2 described in addition, when the correct details mark of placing deposits the file layout that the AFIS system uses in, use then the AFIS matching algorithm for the finger print data library searching of system its.Then, the AFIS system produces the report of matching probability, and if coupling, then this degree of probability is distributed relative mark.On general, the quantity of proper characteristics point mark is high more, and it is big more the probability that AFIS " finds " (i) to occur, and (ii) finds " degree of confidence (confidencelevel) " mark of issuing high more by the AFIS system to each.

Additionally maintain system, equipment, method etc. at this AFIS and can be used in various configurations, comprise following configuration (this catalogue is exemplary and non-exclusive)

1, at " the pre-AFIS " of existing AFIS system level increase/use 3-D visualization tool.This helps the overlooker to check 3-D fingerprint image (for example, the z axle on surface is represented grey measurement level), draws AFIS details mark on the 3D surface, then the details mark is submitted to the AFIS system to be used for search.

2, as a modification of above 1, the 3-D visualization tool also helps overlooker (a) to increase other details after the AFIS system produces details automatically, and (b) adjusts the details of being placed by the automatic placement feature of AFIS system.

3, for the Grade III details that is included in the marking, in this gray scale and the placement of the visual participation details of 3-D of other magnitude.

3-D visualization tool also be applied to keep the seal automatic recognition system (Automated PalmprintIdentifiation Systems) and other recognition system based on the marking, shown in the example among Fig. 6, in Fig. 6, check for more detailed 3D, in the embodiment of this system and method for discussing, the 2D that shows general (that is, with 2D) and draw with the 3D that has the grey measurement level that illustrates in another dimension keeps the seal 38.Keep the seal and fingerprint all comprise the friction wrinkle ridge skin (friction ridge skin) and in a similar manner the search.The marking of other general human expectation for example toe seal, footmark, the non-wrinkle ridge skin marking etc. also can be by imaging, and uses the method and system at this.

Fig. 7-11 has shown the screenshot capture of the embodiment of AFIS 3D details marker software, and shows the exemplary step of being taked for mark wrinkle ridge destination node and bifurcation (grades II details).

Fig. 7 is the screenshot capture that is used in the fingerprint that the embodiment of this 3D details marking tool of discussing draws.Can be in Fig. 7 (with in some other accompanying drawings) see, in certain embodiments, the layout that details is marked on the 3D rendering can be simultaneously displayed on the 2D image, if expectation, as shown, the 2D image can be close to 3D rendering and insert or insert in it, and perhaps image walks abreast, and perhaps is provided with by expectation in addition.

Next, Fig. 8 has described to stand the details mark image 38 of keeping the seal that handle, identical with Fig. 6.Potential the keeping the seal of 2D that first graphical representation is manifested in details mapping software (minutia plotting softward).When details was placed in the 3D rendering, it appeared at the corresponding location on the 2D image of the left side automatically.

Fig. 9 has described the keep the seal image 38 identical with Fig. 6, has shown the image that the rotation that stands along any direction, deflection (skew), rolling etc. are checked with the 3 deep dimensions of impelling wrinkle ridge structure 44.

Figure 10 has described the keep the seal image 38 identical with Fig. 6, and for example the contrast adjustment is observed by the enhancement mode filtrator for it, with the image 45 that filtration is provided.Such filtrator can make in any desired time that to be used for further increasing vision clear.

Figure 11 has described the keep the seal image 38 identical with Fig. 6, and wherein the user also can be placed on details mark 46 on the 3D surface, and it is quantized and searches for for existing 2D database.

Therefore, in certain embodiments, comprise the analysis marking in this method and related software and other system, it comprises: the marking images that at least 2 dimensions a) are provided; B) make this image stand magnitude and strengthen analysis, thereby at least one relative measurement at least one essential part of the marking can be depicted in at least 2 another dimensions of tieing up, with the image that provides magnitude to strengthen, make and do not tie up images, but other magnitude rank can be for human eye more identification basically than there being magnitude to strengthen 2 of analysis; C) show the image that strengthens; And d) image of manual examination (check) magnitude enhancing is to be placed at least one details mark on the marking.

This put procedure can comprise identification and place before on the marking Unidentified at least one, two or more details marks, and can comprise move previous incorrect be placed on the marking at least one, two or more details marks.

This method also can comprise the analysis marking, and it comprises: the marking image of at least 2 dimensions a) is provided, and it comprises by automatic details labeling algorithm determines the details mark, so that automatic details mark to be provided; B) make this image stand magnitude and strengthen analysis, thereby at least one the relative magnitude at least one essential part of the marking can be depicted in another dimension at least 2 dimensions, with the image that provides magnitude to strengthen, make that in addition selected measurement or magnitude rank can be for human eye as seen than there not being magnitude to strengthen 2 dimension images of analysis; C) show the image that strengthens; And d) image of manual examination (check) magnitude enhancing is to assess the correctness of automatic details mark.This method can further comprise at least one details mark of placing the marking, this put procedure comprises that at least one removes incorrect, automatic details mark, move incorrect, automatic details mark, or increase other details mark.

AFIS III+ method and system

When the specific gray scale in connecting fingerprint image or other magnitude, the path of last gained meets specific Grade III edge and pore wrinkle ridge feature.Use many gray scales or other magnitude path to strengthen more basically sometimes Grade III Feature Recognition and extractions that have in the impression more, therefore increased the possibility that AFIS finds.

Have the pixel of same grayscale or other magnitude by connection, form path or profile.This path meets the unique wrinkle ridge shape that occurs along edge, pore and the configuration of surface of friction wrinkle ridge impression.The change of the figure route of figure route (charted course) and general x-y axle path can be used for identification and extracts the Grade III feature.

When selecting different gray-scale values when (or other is used for the value of measured magnitude), path or profile present new route.The change of figure variation route is different with the change of any other path route.In brief, when the magnitude that is used to draw path figure changes, along the shape of this via features, position, outstanding and exist feature also to change.If use the multiple order of magnitude path (at the given measurement indicator multi-path in the gray scale for example at AFIS III+ environment, and/or different magnitude indicators for example in gray scale and form and aspect and the saturation degree or the multi-path between it), can discern and be present in many in the impression and may whole substantially features.

Grades II AFIS (AFIS II) pattern is mainly considered friction wrinkle ridge lines deviation main in the impression.These comprise bifurcation 48 and wrinkle ridge destination node 50 shown in Figure 12 a-c.Figure 12 a (left side) has described the image of known impression, and Figure 12 b (central authorities) illustrates wrinkle ridge lines and the main wrinkle ridge lines deviation that the AFIS system reads, and discerns and extract the grades II details with directivity and relation at Figure 12 c (right side) so.

Grade III AFIS (AFIS III) considers in addition the feature along less important deviation edge, for example along the wrinkle ridge position 52,54 and the pore locations 56 of wrinkle ridge central authorities, if obtainable words are like this in the impression of the marking.Figure 13 a (left side) has described the figure of known impression, and Figure 13 b (central authorities) illustrates AFIS II system and reads wrinkle ridge lines and main wrinkle ridge lines deviation, discerns and extract the grades II details with directivity and relation at Figure 13 c (right side) so.

AFIS III+ considers this details and the more details of the multi-level or impression section of impression as shown in figure 14.These sections are by the gray-scale value definition, and this gray-scale value is used to draw the route of interior profile of section or path.Therefore, Figure 14, AFIS III+ use the section comprise many images path, and it differently meets the Grade III feature in different gray scales (or other) magnitude.

Shown in Figure 15 a-c and Figure 16 a-c, these AFIS III+ paths can be isolated and individually be checked to show the uniqueness of the friction wrinkle ridge that produces impression.The AFIS III+ that Figure 15 a-c has described at the grey measurement level of difference, identical with Figure 12 a-c image analyzes; Figure 16 a-c has described identical but has analyzed with the AFIS III+ of the image of the different grey measurement levels of the analysis of Figure 15 a-c.By checking the route and the change of path route particularly of path, can be at the specific characteristic (Figure 15 b and 16b) of this route of difference section identification.By the change apportioning cost to the route of path, those specific characteristics can be extracted and make and be used for searching for (Figure 15 c and 16c).Therefore, in Figure 15 a and 16a, view each different paths.In Figure 15 b and 16b, the feature of the path that identification is filled (or section).In Figure 15 c and 16c, only shown marker characteristic (relation, direction and outstanding can be relevant) with each feature.

Therefore, if draw path or rank in the marking according to different gray scales or other magnitude, so along the position of the feature of path, place, outstanding and exist and change.

By drawing path based on comprehensive one group of gray-scale value in the image, so a large number of details exists can be recognized, be extracted and be used in the comparison and identification of fingerprint image.AFIS III+ causes the more accurate and complete profile based on potential marking feature (profile), but this profile can need more processing time and energy to search for.This is because the data that the characteristic set of last gained comprises are the data manyfold that traditional rank 2AFIS system acquisition is had.In addition, these data of accelerating can appear in each section of each image of being compared, and each section can compare with a plurality of sections in given data storehouse, and it is shown in Figure 17.

Therefore, in certain embodiments, only use selected section, maybe can provide program to come packed data or impel data storage, management, processing, analysis etc. in addition.

Forward in this more general discussion in this respect of innovation, a feature relates to the use of a plurality of sections of single image under the AFIS environment.Be included in the interior path that uses by gray scale or the definition of other magnitude of image of friction wrinkle ridge impression on the other hand.In case determine the path in each section, can use any extraction or matching algorithm to collect and comparing data so.Other aspect comprises the feature that definition is used to discern.

Four kinds of Grade III features of the direction change that relates to grey measurement level path are generally arranged, and it can occur under AFIS III+ environment and quantize.

1) EC: the peak excursion point of recessed edge feature

2) EV: the peak excursion point of raised brim feature

3) MC: the barycenter of recessed morphological feature

4) MV: the barycenter of projection morphological feature

Morphological feature is such feature, and wherein outline line or path are around the contour of Grade III feature.Figure 18 illustrates the example of recessed morphological feature, its can comprise the depression at pore, wrinkle ridge top or details than around dark pixel have other morphological feature of brighter shades of gray.The example of projection morphological feature comprises the protuberance on beginning wrinkle ridge (incipient ridge) or the friction wrinkle ridge, and wherein bright details demonstrates darker shades of gray around the details ratio.In these accompanying drawings, recessed morphological feature has brighter gray-scale value, and protruding morphological feature has than the dark gray-scale value of details on every side.

As shown in figure 19, the path of edge feature for below friction wrinkle ridge edge, advancing along path, its change by path direction is represented.The example of recessed edge feature is included in highlights inlet (groove (furrow)) details of friction wrinkle ridge or pore, and it not exclusively seals at an edge.The example of raised edges feature is included in the protuberance on a wrinkle ridge sidepiece or the wrinkle ridge part, and it is charged in the groove.In Figure 19, the purple feature is the example of recessed edge feature, and green characteristic is the example of raised brim feature.

Identification

By simple algebraically or other desired mode, can determine and draw the place and the direction of each feature with respect to the further feature of pattern center and x-y axle.

Under the situation of morphological feature, the pixel ash measurement level based on respect to surrounding values can calculate barycenter and characteristic area, but and assign direction.For example, protruding features is accepted "+" value, and protruding features is accepted "-" value.Can (can ignore slight change, threshold values tolerance limit (threshold tolerance) maybe can be set by the outstanding reduction noise of analyzing whole many slice feature.Also can change and reduce noise (false feature is represented in violent variation, and it is not a friction wrinkle ridge skin characteristic) by the number percent of the gray scale pixel magnitude in the pixel around analyzing.

For edge feature, based on respect to the pixel of the value on the arbitrary sidepiece of path ash measurement level, can calculate away from the point on the path of average path, but and assign direction.For example, protruding features is accepted "+" value, and recessed feature is accepted "-" value.Can reduce noise (can ignore slight change, the threshold values of tolerance limit maybe can be set) from the deviation of average path by analyzing this point.Also can reduce noise (frequecy characteristic is represented false feature, and it is not a friction wrinkle ridge skin characteristic) by analyzing along the frequency of the feature of path.

Forward some general discussion to, may increase the accuracy of automatic system of fingerprint recognition significantly in this development of innovation, and/or increased more multinational domestic criminal's identification outward, therefore help the progress of enforcement, criminal justice system and home guard work.

In fact any dimension, at least the weighted array of the dimension of 2D digital picture (for example, Direct Digital image, scanned photograph, obtain from the screen of video or other moving image) can be expressed as 3D exterior view at least (be the dimension of pixel or intensity (or the magnitude of determining by some other mathematical notations or the correlativity of pixel, for example pixel intensity and around it pixel on average or only of pixel intensity average) can be expressed as at least one other dimension; X, y image can be used for producing x, y, z face, and wherein the definition of z axle produces the selected magnitude of z axle).For example, magnitude can be gray scale or given color channel (color channel).

Other example comprises that the acquiescence color space (default color space) of image converts HLS (form and aspect, brightness, the saturation degree) color space to, selects saturation degree, colourity or brightness dimension as magnitude then.Convert the selection (red channel, green channel, blue channel etc.) that the RGB color space allows color channel to.This selection also can be single wavelength or wave band, or is multi-wavelength or wave band.For example, select and/cancellation selects some wave band can allow the fluoroscopic examination of image, or the relative oxygen content of hemochrome in the detected image.Use the mathematical function of linearity or nonlinear algorithm or other expectation can determine magnitude.

Therefore, from the combination of color space dimension (passage) and some weighting factor (for example 0.5 ^*Red+0.25 ^*Green+0.25 ^*Blue), or even simultaneously from the combination (for example, pixel intensity (from the HIS color space) multiplies each other with its brightness (from color spaces such as YUV, YcbCr, Yxy, LAB)) of the dimension of the different color spaces, but the height of each pixel on the gauging surface.

In certain embodiments, pixel meets pixel surface projection and connects by image processing techniques, to produce continuous exterior view.The image processing techniques that is used to connect projection and produces the surface comprises: the 2D pixel is mapped to the lattice point (for example triangle or straight line) of 3D grid, and the z axle value of lattice point is set to suitable value (based on selected tolerance intensity for example, red channel waits to be increased); Fill grid with standard 3D shade technology (Gao Luode, plane etc.); Then with around and the directional lighting device illuminate the 3D scene.Can implement these technology to the embodiment that uses the modification in a certain 3D surface creation/visual software, for example in U.S. Patent No. 6,445,820 and No.6,654,490; U.S. Patent application 20020114508; 20020176619; 20040096098; 20040109608; With discussion among the PCT public announcement of a patent application No.WO 02/17232.

The present invention can show 3D topomap or 1 bit or more other 3D of the color space dimension in the image of higher bit show.For example, 4,096 variations of the variation of the form and aspect in 12 bit image usable surface height are expressed as the 3D surface.

Other example that magnitude and/or demonstration are selected comprises outside the color space dimension that any function of use 2D data set can calculate the lattice point height on the z axle.Being used for changing over from the information of 2D data set the z function form of expression highly is f (x, y, image)=z.All color space dimensions all are this forms, yet also other value can be arranged.For example, can in Lumen software, create function, it is based on the look-up table of (i) Huo Sifeierdeshi unit (Hounsfield unit) (f (pixel value)=Huo Sifeierdeshi value), (ii) only (for example based on the 2D coordinate, f (x, y) 2x+y), (iii) outside storable any other field variable of image, or the (iv) area operator (areaoperator) in the 2D image, for example Gaussian Blur value or Sobel edge detector value are shone upon the z height.

In all cases, external function or data set are relevant with image with some meaningful ways.Software at this can comprise function g, and it is mapped to a certain other external variable (for example, Huo Sifeierdeshi unit) with the pixel in the 2D image, and this value is then as z value (having optional adjustment) highly.Net result is the 3D topomap that is included in the Huo Sifeierdeshi unit in the 2D image; This 3D figure be projected in the 2D image originally on one's body.

Therefore, for example, magnitude can be gray scale, form and aspect, brightness or saturation degree at least one or more a plurality of, perhaps magnitude can comprise from the combination that at least one obtained of gray scale, form and aspect, brightness or saturation degree and concentrate on pixel in the image, by the average of area operator definition.Use linearity or nonlinear function can determine magnitude.

This method also can comprise AFIS II type, AFIS III type and/or AFIS III+ type one or more of mark in the carries out image.

At least 2 dimension images of the marking can be 2 dimension images, and can be the third dimensions with respect to another dimensions of 2 dimensions, and so that the 3D rendering with 3 dimension faces to be provided, it has the 2 dimension bidimensionals of representing with x, y axle of images and the third dimension of representing with the z axle.

This magnitude Analysis can be distinguished the Grade III feature of enough other values of level with the difference marking.It can be that dynamic magnitude strengthens analysis that this magnitude strengthens analysis, and can comprise rolling, tilts and/or rock image.This performance analysis also can comprise performance analysis is added cinema loop, and its expression video or other moving image, tilt or rock and repeated back and forth at wherein specific rolling; Other selection that is used for the film loop comprises: the aspect ratio on surface is changed into some other numbers (plus or minus) from 0, (for example change lighting parameter, mix with the % of direction brightness all around), be applied to angle (making lamp " inswept (sweeping) " from the teeth outwards) lamp of the directional lighting on surface.Image can be digital picture, photographic image, color image or black, white image.The marking can be fingerprint, keep the seal, the local marking, the potential marking.

This method further can comprise, when placing the details mark of the marking, shows the details mark simultaneously on the 2D of marking image and 3D rendering.The 2D image and the 3D rendering of the marking can be simultaneously displayed on the single display screen.

Innovation at this also comprises: computer-implemented program design (computer-implementedprogramming), and it carries out the automatic element of this method; With the computing machine that comprises this computer-implemented program.This computing machine can comprise the distributed network that links computing machine, and for example hand-hold wireless computing machine, and this method can be realized with the hand-hold wireless computing machine.This innovation also comprises the AFIS system, and it comprises computing machine and/or carries out computer-implemented program in this method.

In certain embodiments, remain on the lasting record or the command history of each order that the legacy data collection carries out at this view data model program.In order to turn back to the observation of first front surface object,, or for example hit " cancelling/reform " command key or other desired mode for several times by other mechanism and select corresponding order from the command history drop-down list.When the storage magnitude strengthens image, command history and in history current location storage be parts of images file layout or other complicated link form.The such storage of command history can be an automatic or manual, and can be compulsory or optional.For example, under the situation of the chain of custody (with the test of sample) of expecting court's history and/or sample, for example in the review of law court's contentious procedure declarative evidence, command history can remain pressure (promptly, can not be closed by the user), automated characterization, it writes down each image control and reviews after with the lawyer that is used to oppose or expert or other authority.Similarly, in company's occasion of needs tracking employee behavior, command history can be compulsory.In other occasion, wherein susceptible of proof history is not crucial or expectation, and the command history function can be closed maybe and can wipe.The feature of the embodiment of expectation is that command history is from impossible " vacation ".

In certain embodiments, no matter when open the data pattern file of being stored, the command history of being stored can occur automatically, and can use with identical sequence order.Therefore, command history is " lasting ", and this is that this display file is being stored in the data pattern file behind the input command because it has kept the part of each display file.

In case memory command history, history separately historical or that have copying image product can be posted to another user, and this user imports history then, or access history, and goes up the command set that use is stored at identical primary image (underlying image).This has produced the same demonstration, and actually needn't send this demonstration back and forth.For example want external experts to use the iconic model system to check image in the crime laboratory, but want inside to carry out the demonstration of being drawn or can not transmit easily or receive under the situation of very big computer documents (or a series of this file) at last, it may be favourable.It has also increased the ability that second user crosschecks the initial employed method of user.Therefore, this image can be checked by a user, with the magnitude Analysis that realizes expectation, vertical partially, roll and driftage and/or other show and be provided with, the enhancing image that produces expectation shows that the command history that for example will provide the enhancing image of this expectation to show by Email sends second user to then.It can produce identical primary image then in their computing machine institute installed software, or inserts or duplicate the command history that sends to them in addition, and produces the drafting of former state.Also can make other drafting then, and it is sent it back initial user or arrives other desirable users.In certain embodiments, first user is a force at the core, for example is the supplier of iconic model software, and it has the special speciality of the inspection of image or image type.In other embodiments, the user can be that a plurality of different crime laboratories, medical laboratory, virologist or other have peculiar speciality field but the not specific user who is bound by iconic model software.As desired, the user who also has other.

In another aspect, can advantageously be presented at this marking of discussing and other image, making can be at 8 bit or still less show to have 9 bits that are used for each pixel channel information or the image of more bit magnitudes on the display system of bit.The instrument in this method and system etc. of can being included in comprise surface/wire frame/profile/map lattice, contour interval control, height above sea level ratio and convergent-divergent, false colour/grey scale mapping, color/transparent mapped, surface direction, surface projection's perspective, feature and distant view, comparison form ceramic tile filter and synchronously, image registration, image clone, the color pinup picture contrast control by histogram equalization and range of linearity mapping.These systems etc. are transformed into gray level image intensity, other magnitude the 3D surface expression of magnitude.This transformation causes the basic variation of HVS perception mechanism, and wherein pitch value changes " height above sea level " shape and the form of the selected magnitude correspondence that is formed in corresponding pixel.Height above sea level shape and form can be represented with any selected contrast level or form and aspect, have avoided the gray scale tone to show and the HVS perception problems.Can use the multiple interactive tool that is used for the quantity sense and auxiliary, for example visual no-load voltage ratio (zoom), tilt, rock, the mouse gestures of rotation, applied color value, isoline, linear scale, spatial calibration and characteristics of image measures.

This system etc. is provided on the conventional display device and shows more than 8 bits (more than 256 tones) and more (for example, 16 bits, 65,536 tones) gray tone generally can be distinguished maximum 8 bit gradation.In certain embodiments, this is to shine upon digital magnitude image space information by X and Y-axis at image, draws in Z axle or elevation dimension simultaneously that gray-scale value realizes.The three-dimensional surface of last gained can distribute any desired length and zoom factor to the Z axle, and therefore the demonstration of the half-tone information of the common 256 gray scales restriction that equals or exceeds printer, display and human visual system is provided.

In addition, can be (promptly at the subclass of whole bit sets of " amplification form " display message, fully not compression or at least than the less compression of the remainder of magnitude information), thereby only some section quilt of information shows fully, and remainder is compressed or even " leaving " display screen.For example " form " can be the subclass of whole tonal ranges, for example 256 in 4096.Can locate this " form " and arrive 4096 gray-scale value to check middle pitch " rank " (1920 to 2176), extremely dark " rank " (0 to 255) or other 12 bit levels.For extremely dark example, adjust from 4096 or the 256 gray scales parts (form) of extremely dark (rank) gray scale of other higher bit rank image to show those dark gray, it is to use the easy visible middle pitch rank gray scale to the HVS on the common display device.The demonstration that is equilibrated at of 3840 gray scales (4096 deduct 256) in 12 bit image is gone up invisible.Use optional 3 dimension surfaces, the middle pitch of utmost point low key tone and grey and incandescent tone need not be adjusted (form and rank) just as seen.As the object on 3D surface, for the HVS perception can be quickly (if expectation, or more) obtain all 4096 gray-scale values.

In addition, some surface creation technology of discussing with other places more than this, lasting command history, demonstration selection, software etc. itself are formed in this innovation, comprise the purpose except AFIS analyzes.For example, this technology etc. can be used for situations such as medical science, industry, tooth, court, quality assurance, individual identification.

Be included in those terms that mask body is discussed under this part at these employed all terms and use, unless context or definition are clearly pointed out in addition according to its common meaning.In addition, unless refer else, except dropping in the claim, " or " use comprise " with ", vice versa.Non-limiting term should not be construed to restriction, unless stated otherwise or context clearly point out (for example, " comprise ", " having " and " comprising " general expression " comprise and unrestricted ") in addition.Comprise in the claims singulative for example " one (a) ", " one (an) " and " described (the) " comprise more than a reference, unless stated otherwise or context clearly point out in addition.

" computing machine " is a kind of equipment, and it can the gated sweep instrument, digital image analyzer or processor etc. or at other element of this method and apparatus of discussing.For example, computer-controllable is built in this AFIS that discusses analysis, software, and it determines gray scale or other brightness and/or intensity section etc.On general, computing machine comprises CPU (central processing unit) (CPU) or other logic realization equipment, for example stand-alone computer such as desktop type or laptop computer, have the computing machine of periphery, hand-held, LAN (Local Area Network) or the Internet etc.Computing machine is well-known, and in view of the disclosure, drops in technician's the scope for particular aspects or the desired computing machine of feature selecting.

The scope of native system and method etc. comprises that device adds function, and step adds concept of function.Yet, the term of being explained should not be interpreted into the expression relation of " device adds function " in the claims in this application, unless word " device " is quoted in the claims especially, and under the situation that word " device " is quoted in the claims especially, it is interpreted into the expression relation of " device adds function " in the claims.Similarly, the term of being explained should not be interpreted into the expression relation of " step adds function " in method or process claim in this application, unless word " step " is quoted in the claims especially, and under the situation that word " step " is quoted in the claims especially, it is interpreted into the expression relation of " step adds function " in the claims.

From aforementioned, should be appreciated that, although for the purpose that illustrates, specific embodiment has been described at this, yet under the discussion and the situation of the spirit and scope of claim that do not break away from this, can do various modifications.

Claims (75)

1, a kind of method of analyzing the marking, it comprises:

A) provide the marking at least the two dimension image;

B) make described image stand magnitude and strengthen analysis, thereby at least one the relevant magnitude at least one essential part of the described marking is depicted in respect to described another two-dimentional at least dimension, to provide magnitude to strengthen image, thereby than there not being described magnitude to strengthen the described two dimensional image of analyzing, additional magnitude rank more can be discerned basically for human eye;

C) show described enhancing image; And

D) the described magnitude of hand inspection strengthens image, to place at least one details mark of the described marking.

2, the method for claim 1, wherein said put procedure comprise identification and place previous at least two unrecognized on described marking details marks.

3, the method for claim 1, wherein said put procedure comprise mobile at least two details marks that before are placed on improperly on the described marking.

4, a kind of method of analyzing the marking, it comprises:

A) provide the image of two dimension at least of the marking, described image comprises the details mark of being determined by the automatic details labeling algorithm that automatic details mark is provided;

B) make described image stand magnitude and strengthen analysis, thereby at least one the relevant magnitude at least one essential part of the described marking is depicted in respect to described another two-dimentional at least dimension, to provide magnitude to strengthen image, thereby than there not being described magnitude to strengthen the described two dimensional image of analyzing, additional magnitude rank for human eye as seen;

C) show described enhancing image; And

D) the described magnitude of hand inspection strengthens image, to assess the correctness of described automatic details mark.

5, method as claimed in claim 4, wherein said method further comprises at least one details mark of determining the described marking, and described deterministic process comprises removes incorrect automatic details mark, moves incorrect automatic details mark or increase in the other details mark at least one.

6, as each described method among the claim 1-5, wherein said magnitude is a gray scale.

7, as each described method among the claim 1-5, wherein said magnitude comprises at least one of form and aspect, brightness or saturation degree.

8, as each described method among the claim 1-5, wherein said magnitude comprises from the combination of the value that at least one obtained of gray scale, form and aspect, brightness or saturation degree.

9, as each described method among the claim 1-5, wherein said magnitude comprises by the defined mean intensity of area operator that concentrates on the pixel in the described image.

10, as each described method among the claim 1-5, wherein use linear function to determine described magnitude.

11, as each described method among the claim 1-5, wherein use nonlinear function to determine described magnitude.

12, as each described method among the claim 1-11, wherein said method further comprises the automatic system of fingerprint recognition II type analysis of carrying out the described mark in the described image.

13, as each described method among the claim 1-11, wherein said method further comprises the automatic system of fingerprint recognition III type analysis of carrying out the described mark in the described image.

14, as each described method among the claim 1-13, wherein the described image of two dimension at least of the marking is a two dimensional image, and another dimension with respect to described two dimension is the third dimension, so that the 3-D view with three-dimensional surface to be provided, two x of Wesy of wherein said two dimensional image, y axle are represented, and the described third dimension is represented with the z axle.

15, as each described method among the claim 1-14, wherein said magnitude Analysis is distinguished enough described magnitude ranks to distinguish the Grade III feature of the described marking.

16, as each described method among the claim 1-15, it is that dynamic magnitude strengthens analysis that wherein said magnitude strengthens analysis.

17, method as claimed in claim 16, wherein said performance analysis comprise at least rolls, tilts or rock described image.

18, method as claimed in claim 17, wherein said performance analysis comprise at least rolls, tilts and rock described image.

19, as claim 17 or 18 described methods, wherein said performance analysis comprises makes described performance analysis add cinema loop.

20, as each described method among the claim 1-19, wherein said image is a digital picture.

21, as each described method among the claim 1-19, wherein said image is the digital scanning of photography video

22, as each described method among the claim 1-21, wherein said image is a color image.

23, as each described method among the claim 1-22, wherein said image is black-white image.

24, as each described method among the claim 1-23, the wherein said fingerprint that is imprinted as.

25, as each described method among the claim 1-23, wherein said being imprinted as kept the seal.

26,, wherein saidly be imprinted as the local marking as each described method among the claim 1-25.

27, as each described method among the claim 1-25, the wherein said marking is the potential marking.

28,, when wherein said method further is included in the described details mark of placing the described marking, on the two dimensional image of the described marking and 3-D view, show described details mark simultaneously as each described method among the claim 1-27.

29, method as claimed in claim 28, the described two dimensional image and the described 3-D view of the wherein said marking are simultaneously displayed on the single display screen.

30, enforcement of rights requires the computer-implemented program design of the automatic element of each described method among the 1-29.

31, a kind of computing machine that comprises computer-implemented program design, its enforcement of rights requires the automatic element of each described method among the 1-29.

32, computing machine as claimed in claim 31, wherein said computing machine comprises the distributed network of the computing machine of phase chain.

33, computing machine as claimed in claim 32, wherein said computing machine comprises handheld computer, and each described method realizes with described hand-hold wireless computing machine among the claim 1-29.

34, computing machine as claimed in claim 32, wherein said computing machine comprises the wireless connections computing machine, and each described method realizes with described hand-hold wireless computing machine among the claim 1-29.

35, a kind of system that comprises the automatic system of fingerprint recognition of computer-implemented program design, described computer-implemented program design enforcement of rights requires the automatic element of each described method among the 1-29.

36, the system of automatic system of fingerprint recognition as claimed in claim 32, wherein, the system of described automatic system of fingerprint recognition comprises the hand-hold wireless computing machine, and each described method realizes with described hand-hold wireless computing machine among the claim 1-29.

37, a kind of system that comprises according to the automatic system of fingerprint recognition of each described computing machine among the claim 31-35.

38, a kind of method of analyzing the marking, it comprises:

A) provide the marking at least the two dimension image;

B) make described image stand magnitude and strengthen analysis, thereby at least one the relevant magnitude at least one essential part of the described marking is depicted in respect to another two-dimentional at least dimension, to provide magnitude to strengthen image, thereby than there not being described magnitude to strengthen the two dimensional image of analyzing, additional magnitude rank for human eye as seen;

C) described magnitude is strengthened image and be divided into a plurality of intensity ranks;

D) individually select at least one isolated intensity rank;

E) determine the mark of at least one automatic system of fingerprint recognition from described isolated intensity rank.

39, method as claimed in claim 38, wherein said method further comprise the described isolated rank of demonstration, and manually determine at least one details mark.

40, method as claimed in claim 38 is wherein determined described details mark automatically

41, as each described method among the claim 38-40, wherein said a plurality of intensity ranks comprise at least three ranks.

42, as each described method among the claim 38-41, wherein said method further comprises removes incorrect details mark, moves incorrect details mark or increases in the details mark at least one for described image.

43, as each described method among the claim 38-41, wherein said deterministic process comprises determines previous at least two unrecognized on described marking details marks.

44, as each described method among the claim 38-41, wherein said deterministic process comprises mobile at least two details marks that before are placed on improperly on the described marking.

45, as each described method among the claim 38-44, wherein said intensity level is a gray scale.

46, as each described method among the claim 38-44, wherein said magnitude comprises at least one of form and aspect, brightness or saturation degree.

47, as each described method among the claim 38-44, wherein said magnitude comprises from the combination of two values that obtained of gray scale, form and aspect, brightness or saturation degree at least.

48, as each described method among the claim 38-44, wherein said magnitude comprises by the defined mean intensity of area operator that concentrates on the pixel in the image.

49, as each described method among the claim 38-44, wherein use linear function to determine described magnitude.

50, as each described method among the claim 38-44, wherein use nonlinear function to determine described magnitude.

51, as each described method among the claim 38-44, wherein said method further comprises the automatic system of fingerprint recognition II type analysis of carrying out the mark in the described image.

52, as each described method among the claim 38-44, wherein said method further comprises the automatic system of fingerprint recognition III type analysis of carrying out the mark in the described image.

53, as each described method among the claim 38-44, wherein the described image of two dimension at least of the marking is a two dimensional image, and another dimension with respect to described two dimension is the third dimension, so that the 3-D view with three-dimensional surface to be provided, two x of Wesy of wherein said two dimensional image, y axle are represented, and the described third dimension is represented with the z axle.

54, as each described method among the claim 38-53, wherein said magnitude Analysis is distinguished enough described magnitude ranks to distinguish the Grade III feature of the described marking.

55, as each described method among the claim 38-54, it is that dynamic magnitude strengthens analysis that wherein said magnitude strengthens analysis.

56, method as claimed in claim 55, wherein said performance analysis comprises rolling, tilts or rocks described image.

57, method as claimed in claim 56, wherein said performance analysis comprises rolling, tilts and rocks described image.

58, as claim 56 or 57 described methods, wherein said performance analysis comprises makes described performance analysis add cinema loop.

59, as each described method among the claim 38-58, wherein said image is a digital picture.

60, as each described method among the claim 38-58, wherein said image is the digital scanning of photography video.

61, as each described method among the claim 38-60, wherein said described image is a color image.

62, as each described method among the claim 38-60, wherein said image is a black white image.

63, as each described method among the claim 38-62, the wherein said described fingerprint that is imprinted as.

64, as each described method among the claim 38-62, wherein said described being imprinted as kept the seal.

65,, wherein saidly be imprinted as the local marking as each described method among the claim 38-64.

66, as each described method among the claim 38-64, the wherein said marking is the potential marking.

67,, when wherein said method further is included in the described details mark of placing the described marking, on the two dimensional image of the described marking and 3-D view, show described details mark simultaneously as each described method among the claim 38-66.

68, as the described method of claim 67, the described two dimensional image and the described 3-D view of the wherein said marking are simultaneously displayed on the single display screen.

69, enforcement of rights requires the computer-implemented program design of the automatic element of each described method among the 38-68.

70, a kind of computing machine that comprises computer-implemented program design, its enforcement of rights requires the automatic element of each described method among the 38-68.

71, as the described computing machine of claim 70, wherein said computing machine comprises the distributed network of phase chain computing machine.

72, as the described computing machine of claim 70, wherein said computing machine comprises the hand-hold wireless computing machine, and each described method realizes with described hand-hold wireless computing machine among the claim 38-68.

73, a kind of system that comprises the automatic system of fingerprint recognition of computer-implemented program design, described computer-implemented program design enforcement of rights requires the automatic element of each described method among the 38-68.

74, as the system of the described automatic system of fingerprint recognition of claim 71, the system of wherein said automatic system of fingerprint recognition comprises the hand-hold wireless computing machine, and each described method realizes with described hand-hold wireless computing machine among the claim 38-68.

75, a kind of system that comprises according to the automatic system of fingerprint recognition of each described computing machine among the claim 70-72.

Images