WO2003015029A2 - Dispositif et procede de traitement pour systeme de perception automatique - Google Patents
Dispositif et procede de traitement pour systeme de perception automatique Download PDFInfo
- Publication number
- WO2003015029A2 WO2003015029A2 PCT/FR2002/002849 FR0202849W WO03015029A2 WO 2003015029 A2 WO2003015029 A2 WO 2003015029A2 FR 0202849 W FR0202849 W FR 0202849W WO 03015029 A2 WO03015029 A2 WO 03015029A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- order
- signal
- input
- bus
- reg
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration by the use of histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/28—Indexing scheme for image data processing or generation, in general involving image processing hardware
Definitions
- the invention relates to a processing device for an automatic perception system.
- the device comprises a set of histogram calculation units also called electronic space-time neurons (STN), preferably self-adaptive possibly with anticipation and learning. They are more particularly intended for the perception and processing of images.
- STN space-time neurons
- a method is also object of the invention.
- the selection criteria can be extremely varied, it can be a speed, a shape, a color ... or a combination of these criteria.
- These devices can be used to facilitate the apprehension of a scene or a phenomenon by an observer or to control an automation from the information thus extracted.
- These devices implement a spatial and temporal processing unit which, receiving a video type signal S (PI), produces a certain number of parameters for each pixel. These are for example the speed V, the direction DL, a time constant CO, and a binary validation parameter VL in addition to the delayed video signal VR and the various frame synchronization signals, line and pixel grouped under the name F.
- PI video type signal S
- the object of the invention is to propose an improved automatic perception device compared to this state of the art allowing a more flexible treatment of the responses of the calculation units and their functional and hierarchical grouping while reducing the number of electronic elements in front be implemented compared to the previous solution.
- the invention therefore relates to a processing device for an automatic perception system implementing STN calculation modules receiving data from a data bus and interconnected by a reverse annotation bus.
- the modules are grouped into hierarchical sets comprising the set of order 0 formed of a single module, the set of order 1 formed of the union of several sets of order 0, the sets of order P greater than 1 formed from the combination of sets of order P-1 lower, the hierarchical sets of a given order P sharing a feedback bus, and in that the interconnection of the feedback buses between a order P lower and a higher P + 1 order is carried out via an interconnection module according to an interconnection configuration as a function of selection parameters.
- the interconnection module comprises:
- the interconnection module comprises one or more of a second combined term generator in order to send to the retroannotation bus of lower order as many signals as number of second generator,
- the first combined term generator receives at input 1-1 retroannotation signals, the retroannotation signal produced by the second combined term generator being excluded,
- the second combined term generator receives at input S retroannotation signals and the third combined term generator receives at input S-1 retroannotation signals, the signal retroannotation produced by said third combined term generator being excluded,
- the STN calculation module comprising a histogram calculation unit performing a calculation according to a parameter selected in a data selection unit disposed between a data bus and the calculation unit, said parameter being further taken into account by a classification unit intended to produce a decision-making signal, the calculation of the calculation unit is validated by the signal from the second combined term generator, and the first combined term generator is virtual and produces an output signal equal to the signal input, the decision-making signal being sent directly to the third combined term generator, - the combined term generator comprises means for configurable inversion of input signals Eo, Ei ....
- E (R ⁇ + E 0 ). (R ⁇ " + E;) ... (R ⁇ + E;) (Reg o + Reg. + ... Reg ⁇ ) Reg 0 , Reg !, ..., Reg n corresponding to the parameters,
- the parameters are stored in two groups of n + 1 registers, a first group of registers controlling a first input of a series of n + 1 gates OR EXCLUSIVE whose second input receives the corresponding input signal , a second group of registers controlling after inversion a first input of a series of n + 1 OR gates whose second input receives the output signal from the corresponding EXCLUSIVE OR gates, the outputs of the OR gates being combined in an AND gate with n +2 inputs, the AND NOT combination of the inverted outputs of the second group of registers also being sent to the AND gate, to produce an output signal,
- the parameters are stored in two groups of n + 1 registers, a first group of registers controlling a first input of a series of n + 1 gates OR EXCLUSIVE whose second input receives the corresponding input signal , a second register group controlling a first input of a series of n + 1 AND gates, the second input of which receives the output signal of the corresponding EXCLUSIVE OR gates, the outputs of the AND gates being combined in an OR gate with n + 1 inputs to produce an output signal,
- the parameters are stored in two groups of n + 1 registers, a first register group controlling a first input of a first series of n + 1 gates OR EXCLUSIVE whose second input receives the input signal corresponding, a second group of registers controlling after inversion a first input of a series of n + 1 OR gates whose second input receives the output signal from the corresponding EXCLUSIVE OR gates, the outputs of the OR gates being combined in an AND gate with n + 2 inputs, the AND NOT combination of the inverted outputs of the second group of registers also being sent to the AND gate, to produce a first signal and in that for the OR combination, the parameters are stored in two groups of n + 1 registers, a third group of registers controlling a first input of a second series of n + 1 gates OR EXCLUSIVE whose second input receives the corresponding input signal, a fourth g register group controlling a first input of a series of n + 1 AND gates, the second input of which receives
- the first and third registers as well as the first and second series of EXCLUSIVE gates are respectively grouped into a single register and a single series of gates, the outputs of the EXCLUSIVE gates being sent on the one hand to the inputs of the OR gates for the AND combination and on the other hand on the inputs of the AND gates for the OR combination,
- a hierarchical set of modules forms a hierarchical memory of an element of a perception environment and which is defined by the configuration of the parameters for selecting the interconnection modules of the assembly.
- the present invention for order 0 is a processing device for an automatic perception system comprising at least one set of N STN calculation modules, a data bus, a retroannotation bus with N retroannotation signal lines, the modules being interconnected and each of the modules comprising a calculation unit histogram performing a calculation according to a parameter selected in a data selection unit disposed between the data bus and the calculation unit, said parameter being further taken into account by a classification unit intended to produce a decision-making signal according to the result of the calculation unit, the decision-making signal being returned to on the feedback bus, the calculation of the calculation being validated by a signal coming from a retro-annotation selection unit arranged between the retro-annotation bus and the calculation unit, in which each of the modules further comprises a combined term generator receiving on the one hand, a decision-making signal from the classification unit and, on the other hand, N-1 lines of retroannotation signals, the retroannotation signal of
- the invention also relates to a processing method for an automatic perception system implementing STN calculation modules receiving data from a data bus and interconnected by a reverse annotation bus.
- the modules are grouped into hierarchical sets comprising the set of order 0 formed of a single module, the set of order 1 formed by the union of several sets of order 0, the sets of order P greater than 1 formed of the assembly of sets of order P-1 lower, the hierarchical sets of a given order P sharing a feedback bus , and in that the retroannotation buses are interconnected between a lower P order and a higher P + 1 order by means of an interconnection module according to an interconnection configuration as a function of selection parameters.
- a functional hierarchical structure performing calculations, producing results in the form of decision-making returned on a feedback bus, can be easily achieved.
- This structure is tree-like between the different group levels.
- this (super) module being however complex in the case of P> 1 in that it groups together one, and preferably, several modules.
- the subsets are not disjoint since it is possible that a higher order group is formed from the combination of one, and preferably, several lower order groups. It is thus possible to functionally group together according to parameters, sub-sets of modules to form a (super) functional module and it is also possible to functionally group together according to parameters several of these (super) modules.
- the calculations within the system are therefore hierarchical and within each hierarchical level the number and calculations of modules, (super) modules, etc. depending on the level of hierarchy, can be configurable.
- the invention also makes it possible to optimize the size of the buses compared with a device of the state of the art because with the invention the buses between the various improved modules and / or the super modules of several hierarchical groups convey resulting data configurable combinations unlike the state of the art.
- FIG. 2 is a representation of the application of the invention to an STN module
- FIG. 3 is a representation of the application of the invention to a set of the same order of modules
- FIG. 4 is a representation of a hierarchical application
- FIG. 5 is a symbolic representation of an interconnection module between buses
- FIG. 6 is a symbolic representation of an interconnection module in an STN module
- FIG. 7 is a representation of an exemplary embodiment of a combined term generator of the AND type;
- FIG. 8 is a representation of an embodiment of a combined OR type term generator;
- FIG. 9 is a representation of an exemplary embodiment of a combined term generator of mixed AND and OR type.
- FIG. 1 represents the state of the art with two modules 1 of STN calculations.
- Each module 1 includes a calculation unit 2 histogram performing a calculation according to a parameter selected in a data selection unit 5 disposed between a data bus 7 and the calculation unit 2, said parameter being further taken into account by a classification unit 3 intended to produce at output a decision-making signal 9, the decision-making signal 9 is sent on a retroannotation bus 6.
- the calculation of the calculation unit 2 is validated by a signal 8 coming from a retro-annotation selection unit 4 disposed between the retro-annotation bus 6 and the calculation unit 2.
- the STN modules are interconnected on a retroannotation bus 6.
- FIG. 2 the invention is implemented with an STN module.
- a combined term generator 21 is interposed between the output of the classification unit 3 producing the decision-making signal 9 and the feedback bus 6.
- the combined term generator 21 receives as input, in addition to the decision-making signal 9, part of the signals from the feedback bus 6.
- the combined term generator 21 outputs a signal 22 which is sent to the feedback bus 6.
- the signal portion of the retroannotation bus 6 which is sent to the input of the combined term generator 21 corresponds to all of said bus except the signal line 22, that is to say the line produced by the STN module considered.
- the unit 4 for retro-annotation selection is replaced by a combined term generator 4 ′ to produce the signal 8 used to validate the calculation of the unit 2 of calculation. All of these elements form an improved STN module 20 and corresponds to an order 0 in the hierarchy that the invention allows.
- Figure 3 allows to visualize a set n of improved STN modules, 20o up to 20 n - ⁇ and interconnected on a feedback bus of n + 1 signals, the additional signal 33 being produced by an interconnection module between groups d 'lower and higher order.
- Such a structure corresponds to an order 1, order of the retroannotation bus 6 connecting improved STN modules, in the hierarchy which the invention allows.
- the additional signal 33 on this order 1 reverse annotation bus allows the taking into account of results coming from a higher order reverse annotation bus 36.
- the interconnection module between lower order 6 and higher order retroannotation bus 36 comprises three combined term generators whose structure is equivalent to that implemented in an improved STN module with the combined term generator 32 corresponding to the generator 4 ', the generator 34 to the generator 21 and the generator 30 not having to be in an improved STN module since the STN module produces only one decision-making signal 9.
- the correspondence for the signals is 31 to 9 and 33 to 8 and 37 to 22.
- Figure 4 corresponds to a structure comprising several orders, order 0 for individual improved STN modules, order 1 for each of the p groups 35o to 35 p- ⁇ of improved STN modules grouped on their retroannotation bus and order 2 for all the groups forming a super group 45, associating the P groups 35 0 to 35 p- ⁇ , this super group being itself interconnected to a higher order retroannotation bus 46 by an interconnection module comprising the three combined term generators as previously mentioned.
- the generators for the lower order 30, 32, 34 correspond respectively to 40, 42, 44 for the higher order and the signals 31, 33, 37 to 41, 43, 47. With respect to the order bus higher 46, this hierarchical association of advanced STN modules appears as a super module 45.
- FIG. 1 schematizes in generalized form the interconnection module between a bus Bi of lower order and a bus Bs of higher order.
- the interconnection module 200 comprises:
- a first combined term generator 100a receiving at input at least one lower order bus retroannotation signal and producing at output a decision-making signal as a function of selection parameters
- a second combined term generator 100c receiving as input the higher order retroannotation bus and outputting a signal sent on the lower order retroannotation bus as a function of selection parameters
- a third combined term generator 100b receiving, on the one hand, the higher order retroannotation bus and, on the other hand, the decision-making signal and producing as output a retroannotation signal 300 returned to the higher order reverse annotation bus based on selection parameters.
- more than a second combined term generator 100c is used to send more than one feedback signal to the lower order annotation bus.
- the signal produced by the interconnection module to the bus of higher order is not returned to the input of the third generator 100b of combined term.
- Figure 6 is a representation adapted from Figure 2 in order to show the interconnection module part 200 in an improved STN module.
- the classifier 3 produces only one signal, the first generator 100a of combined term does not have to be and it is therefore represented in dotted lines as being virtual.
- Figures 7 and 8 give two embodiments of AND combinations and OR combinations respectively in the form of modules 100 and 100 '.
- the parameter bus 17 and the input signals 6 are at the inputs of the modules 100 and 100 '.
- FIG. 9 gives an example of a generator 100 "of combined term, the output 109 of which can result from the choice of an AND combination 101 of output 107 or of an OR combination 102 of output 108 by a multiplexer 103 controlled by a signal 19 in function of a parameter from bus 17 and stored in a register 18.
- the combinatorial OR EXCLUSIVE circuit 104 is common to the two circuits 101 and 102.
- a hierarchical set of modules is a memory which can correspond to a particular element which is analyzed in an environment by the automatic perception system.
- the memory can be defined by the configuration of the parameters for selecting the interconnection modules within the assembly. In an enriched environment, that is to say comprising many elements, the device makes it possible to memorize according to the configuration of the selection parameters, by carrying out more functional combinations between bus signals of different orders and therefore between sub-assemblies of modules.
Abstract
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP02772498A EP1419482A2 (fr) | 2001-08-10 | 2002-08-09 | Dispositif et procede de traitement pour systeme de perception automatique |
US10/486,768 US7313551B2 (en) | 2001-08-10 | 2002-08-09 | Processing device and method for an automatic perception system |
JP2003519883A JP2004538583A (ja) | 2001-08-10 | 2002-08-09 | 自動認識システム用の処理装置および方法 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
FR01/10750 | 2001-08-10 | ||
FR0110750A FR2828613B1 (fr) | 2001-08-10 | 2001-08-10 | Dispositif et procede de traitement pour systeme de perception automatique |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2003015029A2 true WO2003015029A2 (fr) | 2003-02-20 |
WO2003015029A3 WO2003015029A3 (fr) | 2003-11-27 |
Family
ID=8866486
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/FR2002/002849 WO2003015029A2 (fr) | 2001-08-10 | 2002-08-09 | Dispositif et procede de traitement pour systeme de perception automatique |
Country Status (5)
Country | Link |
---|---|
US (1) | US7313551B2 (fr) |
EP (1) | EP1419482A2 (fr) |
JP (1) | JP2004538583A (fr) |
FR (1) | FR2828613B1 (fr) |
WO (1) | WO2003015029A2 (fr) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR2828613B1 (fr) * | 2001-08-10 | 2003-11-07 | Holding Bev Sa | Dispositif et procede de traitement pour systeme de perception automatique |
JP5528976B2 (ja) * | 2010-09-30 | 2014-06-25 | 株式会社メガチップス | 画像処理装置 |
Family Cites Families (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR2488420B1 (fr) | 1980-08-07 | 1986-04-04 | Urien Michel | Dispositif d'obtention de l'histogramme des distances dans le temps entre des evenements successifs |
US4783828A (en) | 1986-06-02 | 1988-11-08 | Honeywell Inc. | Two-dimensional object recognition using chain codes, histogram normalization and trellis algorithm |
FR2611063B1 (fr) | 1987-02-13 | 1989-06-16 | Imapply | Procede et dispositif de traitement en temps reel d'un flot de donnees sequence, et application au traitement de signaux video numeriques representatifs d'une image video |
JPH01158579A (ja) | 1987-09-09 | 1989-06-21 | Aisin Seiki Co Ltd | 像認識装置 |
EP0380659A4 (en) | 1987-10-02 | 1992-07-15 | Kabushiki Kaisha Komatsu Seisakusho | Method of recognizing linear pattern |
US5163095A (en) | 1988-04-22 | 1992-11-10 | Toa Medical Electronics Co., Ltd. | Processor for extracting and memorizing cell images, and method of practicing same |
US5109425A (en) | 1988-09-30 | 1992-04-28 | The United States Of America As Represented By The United States National Aeronautics And Space Administration | Method and apparatus for predicting the direction of movement in machine vision |
DE3913620A1 (de) | 1989-04-25 | 1990-10-31 | Fraunhofer Ges Forschung | Verfahren zur bildauswertung |
US5088488A (en) | 1989-12-22 | 1992-02-18 | Medtronic, Inc. | Method and apparatus for implementing histogram storage and trend analysis in a medical stimulator |
JP2834304B2 (ja) * | 1990-10-19 | 1998-12-09 | 松下電器産業株式会社 | 線形電力増幅回路 |
DE69221665T2 (de) | 1991-05-23 | 1997-12-11 | Fuji Photo Film Co Ltd | Verfahren zur Ermittlung einer Belichtung |
KR940011881B1 (ko) | 1991-12-23 | 1994-12-27 | 주식회사 금성사 | 움직임 검출 및 추정장치 |
US5712729A (en) | 1992-04-17 | 1998-01-27 | Olympus Optical Co., Ltd. | Artificial retina cell, artificial retina and artificial visual apparatus |
DE4213717A1 (de) | 1992-04-25 | 1993-10-28 | Asea Brown Boveri | Verfahren zur Ermittlung eines Häufigkeits-Zeitprofils von Ereignissen sowie Vorrichtung zur Durchführung des Verfahrens |
US5384865A (en) | 1992-06-01 | 1995-01-24 | Eastman Kodak Company | Adaptive, hybrid median filter for temporal noise suppression |
JP2940317B2 (ja) | 1992-06-24 | 1999-08-25 | 三菱電機株式会社 | 画像処理装置 |
JPH06205780A (ja) | 1993-01-12 | 1994-07-26 | Aloka Co Ltd | 超音波画像処理装置 |
US5592226A (en) | 1994-01-26 | 1997-01-07 | Btg Usa Inc. | Method and apparatus for video data compression using temporally adaptive motion interpolation |
US5774581A (en) | 1994-02-16 | 1998-06-30 | U.S. Philips Corporation | Device for segmenting a discrete assembly of data |
US5592237A (en) | 1994-11-04 | 1997-01-07 | Infimed, Inc. | High resolution image processor with multiple bus architecture |
US5793888A (en) | 1994-11-14 | 1998-08-11 | Massachusetts Institute Of Technology | Machine learning apparatus and method for image searching |
JPH08292998A (ja) | 1995-04-20 | 1996-11-05 | Mitsubishi Electric Corp | 画像検出装置及び画像検出方法 |
WO2000011610A1 (fr) | 1998-08-25 | 2000-03-02 | Holding Bev S.A. | Appareil et procede pour le traitement d'images |
FR2751772B1 (fr) * | 1996-07-26 | 1998-10-16 | Bev Bureau Etude Vision Soc | Procede et dispositif fonctionnant en temps reel, pour le reperage et la localisation d'une zone en mouvement relatif dans une scene, ainsi que pour la determination de la vitesse et la direction du deplacement |
US7181047B2 (en) * | 1996-07-26 | 2007-02-20 | Patrick Pirim | Methods and apparatus for identifying and localizing an area of relative movement in a scene |
FR2773521B1 (fr) | 1998-01-15 | 2000-03-31 | Carlus Magnus Limited | Procede et dispositif pour surveiller en continu l'etat de vigilance du conducteur d'un vehicule automobile, afin de detecter et prevenir une tendance eventuelle a l'endormissement de celui-ci |
US6115689A (en) | 1998-05-27 | 2000-09-05 | Microsoft Corporation | Scalable audio coder and decoder |
EP0973128B1 (fr) | 1998-06-22 | 2003-06-11 | Texas Instruments Incorporated | Extension d'intensité sélective |
US6597738B1 (en) | 1999-02-01 | 2003-07-22 | Hyundai Curitel, Inc. | Motion descriptor generating apparatus by using accumulated motion histogram and a method therefor |
FR2805629B1 (fr) | 2000-02-24 | 2002-08-30 | Holding Bev Sa | Procede et dispositif de perception automatique |
US7043465B2 (en) * | 2000-02-24 | 2006-05-09 | Holding B.E.V.S.A. | Method and device for perception of an object by its shape, its size and/or its orientation |
US7136842B2 (en) * | 2000-02-24 | 2006-11-14 | Holding B.E.V. S.A. | Method and device for adapting a signal for the detection of the movement of an object |
FR2828613B1 (fr) * | 2001-08-10 | 2003-11-07 | Holding Bev Sa | Dispositif et procede de traitement pour systeme de perception automatique |
WO2005008579A2 (fr) * | 2003-07-16 | 2005-01-27 | Idaho Research Foundation, Inc. | Neurone artificiel biomimetique |
-
2001
- 2001-08-10 FR FR0110750A patent/FR2828613B1/fr not_active Expired - Fee Related
-
2002
- 2002-08-09 EP EP02772498A patent/EP1419482A2/fr not_active Withdrawn
- 2002-08-09 JP JP2003519883A patent/JP2004538583A/ja active Pending
- 2002-08-09 WO PCT/FR2002/002849 patent/WO2003015029A2/fr not_active Application Discontinuation
- 2002-08-09 US US10/486,768 patent/US7313551B2/en not_active Expired - Fee Related
Non-Patent Citations (3)
Title |
---|
FUTAMI R ET AL: "A NEURAL NETWORK MODEL BASED ON SHORT-TERM MEMORIES FOR THE HIERARCHICAL RECOGNITION OF TEMPORAL PATTERNS" SYSTEMS & COMPUTERS IN JAPAN, SCRIPTA TECHNICA JOURNALS. NEW YORK, US, vol. 19, no. 2, 1 février 1988 (1988-02-01), pages 102-109, XP000099242 ISSN: 0882-1666 * |
LIPO WANG: "Spatio-temporal sequence processing with the counterpropagation neural network" SYSTEMS, MAN AND CYBERNETICS, 1996., IEEE INTERNATIONAL CONFERENCE ON BEIJING, CHINA 14-17 OCT. 1996, NEW YORK, NY, USA,IEEE, US, 14 octobre 1996 (1996-10-14), pages 831-835, XP010206910 ISBN: 0-7803-3280-6 * |
MIYAMOTO H ET AL: "Recognition Of Spatio-temporal Patterns By A Multi-layered Neural Network Model" PROCEEDINGS OF1993 JOINT CONFERENCE ON NEURAL NETWORKS, 1993, XP010301692 * |
Also Published As
Publication number | Publication date |
---|---|
JP2004538583A (ja) | 2004-12-24 |
US20050049828A1 (en) | 2005-03-03 |
FR2828613B1 (fr) | 2003-11-07 |
US7313551B2 (en) | 2007-12-25 |
WO2003015029A3 (fr) | 2003-11-27 |
EP1419482A2 (fr) | 2004-05-19 |
FR2828613A1 (fr) | 2003-02-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | Benchmarking single-image dehazing and beyond | |
Jiang et al. | Rain-free and residue hand-in-hand: A progressive coupled network for real-time image deraining | |
Wan et al. | Region-aware reflection removal with unified content and gradient priors | |
TWI766201B (zh) | 活體檢測方法、裝置以及儲存介質 | |
KR20210049947A (ko) | 샘플링된 컬러 정보에 기초하여 이미지 시퀀스들을 생성하고 송신하기 위한 시스템들 및 방법들 | |
EP1259939B1 (fr) | Procede et dispositif de perception automatique | |
FR2674653A1 (fr) | Procede et dispositif de detection de bord pour un systeme de traitement d'image. | |
FR2751772A1 (fr) | Procede et dispositif fonctionnant en temps reel, pour le reperage et la localisation d'une zone en mouvement relatif dans une scene, ainsi que pour la determination de la vitesse et la direction du deplacement | |
CN112598045A (zh) | 训练神经网络的方法、图像识别方法及图像识别装置 | |
CN111079507B (zh) | 一种行为识别方法及装置、计算机装置及可读存储介质 | |
CN111275638A (zh) | 基于多通道注意力选择生成对抗网络的人脸修复方法 | |
US20210365714A1 (en) | Detecting Artificial Multimedia Content Using Deep Network Response Analysis | |
CN107547803A (zh) | 视频分割结果边缘优化处理方法、装置及计算设备 | |
CN116861262B (zh) | 一种感知模型训练方法、装置及电子设备和存储介质 | |
US11188792B2 (en) | Defect detection using multiple models | |
EP1419482A2 (fr) | Dispositif et procede de traitement pour systeme de perception automatique | |
EP3712775A1 (fr) | Procédé et dispositif de détermination de la taille mémoire globale d'une zone mémoire globale allouée aux données d'un réseau de neurones compte tenu de sa topologie | |
EP2473970B1 (fr) | Dispositif de calcul reconfigurable sur voisinage flexible de pixels | |
EP1364341A1 (fr) | Procede et dispositif de reperage d'un objet par sa forme, sa dimension et/ou son orientation | |
EP3616132A1 (fr) | Procédé et dispositif automatisés aptes à assurer l'invariance perceptive d'un évènement spatio-temporel dynamiquement en vue d'en extraire des représentations sémantiques unifiées | |
Chang et al. | Coarse-to-fine Feedback Guidance Based Stereo Image Quality Assessment Considering Dominant Eye Fusion | |
WO2024032966A1 (fr) | Procédé de traitement d'image numérique | |
EP1095358B1 (fr) | Procede de modelisation d'objets ou de scenes 3d | |
CN116310579A (zh) | 车辆损伤检测方法、装置、设备及存储介质 | |
Gortázar | Navigating Subjectivity in AI-Generated Photography: The Quest for Ethics and Creative Creative Agency |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BY BZ CA CH CN CO CR CU CZ DE DM DZ EC EE ES FI GB GD GE GH HR HU ID IL IN IS JP KE KG KP KR LC LK LR LS LT LU LV MA MD MG MN MW MX MZ NO NZ OM PH PL PT RU SD SE SG SI SK SL TJ TM TN TR TZ UA UG US UZ VC VN YU ZA ZM Kind code of ref document: A2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ OM PH PL PT RO RU SD SE SG SI SK SL TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A2 Designated state(s): GH GM KE LS MW MZ SD SL SZ UG ZM ZW AM AZ BY KG KZ RU TJ TM AT BE BG CH CY CZ DK EE ES FI FR GB GR IE IT LU MC PT SE SK TR BF BJ CF CG CI GA GN GQ GW ML MR NE SN TD TG Kind code of ref document: A2 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR IE IT LU MC NL PT SE SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2002772498 Country of ref document: EP Ref document number: 2003519883 Country of ref document: JP |
|
WWP | Wipo information: published in national office |
Ref document number: 2002772498 Country of ref document: EP |
|
REG | Reference to national code |
Ref country code: DE Ref legal event code: 8642 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 10486768 Country of ref document: US |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 2002772498 Country of ref document: EP |