WO2019055114A1 - Système visionique sans prise de vue sensible aux attributs par l'intermédiaire de représentations éparses communes - Google Patents

Système visionique sans prise de vue sensible aux attributs par l'intermédiaire de représentations éparses communes Download PDF

Info

Publication number
WO2019055114A1
WO2019055114A1 PCT/US2018/041806 US2018041806W WO2019055114A1 WO 2019055114 A1 WO2019055114 A1 WO 2019055114A1 US 2018041806 W US2018041806 W US 2018041806W WO 2019055114 A1 WO2019055114 A1 WO 2019055114A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
semantic
set forth
training
input image
Prior art date
Application number
PCT/US2018/041806
Other languages
English (en)
Inventor
Soheil KOLOURI
Mohammad Rostami
Kyungnam Kim
Yuri Owechko
Original Assignee
Hrl Laboratories, Llc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US15/949,896 external-priority patent/US10755149B2/en
Application filed by Hrl Laboratories, Llc filed Critical Hrl Laboratories, Llc
Priority to EP18856563.4A priority Critical patent/EP3682370A4/fr
Priority to CN201880052204.6A priority patent/CN111052144A/zh
Publication of WO2019055114A1 publication Critical patent/WO2019055114A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2134—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on separation criteria, e.g. independent component analysis
    • G06F18/21345—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on separation criteria, e.g. independent component analysis enforcing sparsity or involving a domain transformation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/70—Labelling scene content, e.g. deriving syntactic or semantic representations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/10—Recognition assisted with metadata

Definitions

  • the present invention relates to a system for object recognition and, more particularly, to a system for object recognition via joint sparse representations.
  • Zero-shot learning is being able to perform a task despite not having received any training examples of that task.
  • Zero-shot machine vision methods are described by Akata et al. (see Literature Reference No. 1 of the List of
  • the present invention relates to a system for object recognition and, more particularly, to a system for object recognition via joint sparse representations.
  • the system comprises one or more processors and a memory having instructions such that when the instructions are executed, the one or more processors performs multiple operations.
  • a training image set and annotated semantic attributes are used to train a model that maps visual features from known images to the annotated semantic attributes using joint sparse representations with respect to dictionaries of visual features and semantic attributes.
  • the trained model is used for mapping visual features of an unseen input image to its semantic attributes.
  • the unseen input image is classified as belonging to an image class, and a device is controlled based on the classification of the unseen input image, wherein the device is a vehicle component, and controlling the device results in a vehicle maneuver.
  • the system generates a training image set comprising object images from a plurality of image classes, wherein each object image in the training image set has been annotated with a class label and semantic attributes describing the object image.
  • semantic attribute space are modeled as nonlinear spaces that provide an identical sparse representation for visual features and their corresponding semantic attributes.
  • the system finds a sparse representation for a visual feature extracted from the unseen input image, and a semantic attribute prediction is generated that is resolved in the semantic attribute space of the model, wherein a soft-assignment probability vector identifies a probability of the semantic attribute prediction belonging to a class of unseen images.
  • a regularization parameter is used to regulate entropy of the soft-assignment probability vector.
  • the unseen input image is labeled using a class label of a closest semantic attribute in the semantic attribute space of the model.
  • the vehicle maneuver is a collision avoidance maneuver.
  • the unseen input image is an image of an
  • the present invention also includes a computer program product and a computer implemented method.
  • the computer program product includes computer-readable instructions stored on a non-transitory computer-readable medium that are executable by a computer having one or more processors, such that upon execution of the instructions, the one or more processors perform the operations listed herein.
  • the computer implemented method includes an act of causing a computer to execute such instructions and perform the resulting operations.
  • FIG. 1 is a block diagram depicting the components of a system for object recognition according to some embodiments of the present disclosure
  • FIG. 2 is an illustration of a computer program product according to some embodiments of the present disclosure
  • FIG. 3 is an illustration of the training phase of the zero-shot machine vision system according to some embodiments of the present disclosure
  • FIG. 4 is an illustration of finding a sparse representation of an image
  • FIG. 5 is an illustration of identification of the probability of prediction that an attribute belongs to an unseen class of images according to some
  • FIG. 6 is an illustration of a chart depicting test classification accuracy for a dataset according to some embodiments of the present disclosure.
  • FIG. 7 is an illustration of using a processor to control a device based on the classification of an unseen image according to some embodiments of the present disclosure.
  • the present invention relates to a system for object recognition and, more particularly, to a system for object recognition via joint sparse representations.
  • the following description is presented to enable one of ordinary skill in the art to make and use the invention and to incorporate it in the context of particular applications. Various modifications, as well as a variety of uses in different applications will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to a wide range of aspects. Thus, the present invention is not intended to be limited to the aspects presented, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein. [00034] In the following detailed description, numerous specific details are set forth in order to provide a more thorough understanding of the present invention. However, it will be apparent to one skilled in the art that the present invention may be practiced without necessarily being limited to these specific details. In other instances, well-known structures and devices are shown in block diagram form, rather than in detail, in order to avoid obscuring the present invention.
  • the first is a system for object recognition.
  • the system is typically in the form of a computer system operating software or in the form of a "hard-coded" instruction set. This system may be incorporated into a wide variety of devices that provide different functionalities.
  • the second principal aspect is a method, typically in the form of software, operated using a data processing system (computer).
  • the third principal aspect is a computer program product.
  • the computer program product generally represents computer-readable instructions stored on a non-transitory computer-readable medium such as an optical storage device, e.g., a compact disc (CD) or digital versatile disc (DVD), or a magnetic storage device such as a floppy disk or magnetic tape.
  • a non-transitory computer-readable medium such as an optical storage device, e.g., a compact disc (CD) or digital versatile disc (DVD), or a magnetic storage device such as a floppy disk or magnetic tape.
  • Other, non-limiting examples of computer-readable media include hard disks, read-only memory
  • FIG. 1 A block diagram depicting an example of a system (i.e., computer system
  • the computer system 100 is configured to perform calculations, processes, operations, and/or functions associated with a program or algorithm.
  • certain processes and steps discussed herein are realized as a series of instructions (e.g., software program) that reside within computer readable memory units and are executed by one or more processors of the computer system 100. When executed, the instructions cause the computer system 100 to perform specific actions and exhibit specific behavior, such as described herein.
  • the computer system 100 may include an address/data bus 102 that is
  • processor 104 configured to communicate information. Additionally, one or more data processing units, such as a processor 104 (or processors), are coupled with the address/data bus 102.
  • the processor 104 is configured to process information and instructions.
  • the processor 104 is a microprocessor.
  • the processor 104 may be a different type of processor such as a parallel processor, application-specific integrated circuit (ASIC), programmable logic array (PLA), complex programmable logic device (CPLD), or a field programmable gate array (FPGA).
  • ASIC application-specific integrated circuit
  • PLA programmable logic array
  • CPLD complex programmable logic device
  • FPGA field programmable gate array
  • the computer system 100 is configured to utilize one or more data storage units.
  • the computer system 100 may include a volatile memory unit 106 (e.g., random access memory (“RAM”), static RAM, dynamic RAM, etc.) coupled with the address/data bus 102, wherein a volatile memory unit 106 is configured to store information and instructions for the processor 104.
  • RAM random access memory
  • static RAM static RAM
  • dynamic RAM dynamic RAM
  • the computer system 100 further may include a non-volatile memory unit 108 (e.g., read-only memory (“ROM”), programmable ROM (“PROM”), erasable programmable ROM (“EPROM”), electrically erasable programmable ROM “EEPROM”), flash memory, etc.) coupled with the address/data bus 102, wherein the nonvolatile memory unit 108 is configured to store static information and instructions for the processor 104.
  • the computer system 100 may execute instructions retrieved from an online data storage unit such as in "Cloud” computing.
  • the computer system 100 also may include one or more interfaces, such as an interface 1 10, coupled with the address/data bus 102.
  • the one or more interfaces are configured to enable the computer system 100 to interface with other electronic devices and computer systems.
  • the communication interfaces implemented by the one or more interfaces may include wireline (e.g., serial cables, modems, network adaptors, etc.) and/or wireless (e.g., wireless modems, wireless network adaptors, etc.) communication technology.
  • the computer system 100 may include an input device 112 coupled with the address/data bus 102, wherein the input device 112 is configured to communicate information and command selections to the processor 100.
  • the input device 112 is an alphanumeric input device, such as a keyboard, that may include alphanumeric and/or function keys.
  • the input device 112 may be an input device other than an alphanumeric input device.
  • the computer system 100 may include a cursor control device 114 coupled with the address/data bus 102, wherein the cursor control device 114 is configured to communicate user input information and/or command selections to the processor 100.
  • the cursor control device 114 is implemented using a device such as a mouse, a track-ball, a track-pad, an optical tracking device, or a touch screen.
  • the cursor control device 114 is directed and/or activated via input from the input device 112, such as in response to the use of special keys and key sequence commands associated with the input device 112.
  • the cursor control device 114 is configured to be directed or guided by voice commands.
  • the computer system 100 further may include one or more optional computer usable data storage devices, such as a storage device 116, coupled with the address/data bus 102.
  • the storage device 116 is configured to store information and/or computer executable instructions.
  • the storage device 116 is a storage device such as a magnetic or optical disk drive (e.g., hard disk drive (“HDD”), floppy diskette, compact disk read only memory (“CD-ROM”), digital versatile disk (“DVD”)).
  • a display device 118 is coupled with the address/data bus 102, wherein the display device 118 is configured to display video and/or graphics.
  • the display device 118 may include a cathode ray tube (“CRT”), liquid crystal display (“LCD”), field emission display (“FED”), plasma display, or any other display device suitable for displaying video and/or graphic images and alphanumeric characters recognizable to a user.
  • CTR cathode ray tube
  • LCD liquid crystal display
  • FED field emission display
  • plasma display or any other display device suitable for displaying video and/or graphic images and alphanumeric characters recognizable to a user.
  • the computer system 100 presented herein is an example computing
  • the non-limiting example of the computer system 100 is not strictly limited to being a computer system.
  • an aspect provides that the computer system 100 represents a type of data processing analysis that may be used in accordance with various aspects described herein.
  • other computing systems may also be
  • one or more operations of various aspects of the present technology are controlled or implemented using computer-executable instructions, such as program modules, being executed by a computer.
  • program modules include routines, programs, objects, components and/or data structures that are configured to perform particular tasks or implement particular abstract data types.
  • an aspect provides that one or more aspects of the present technology are implemented by utilizing one or more distributed computing environments, such as where tasks are performed by remote processing devices that are linked through a communications network, or such as where various program modules are located in both local and remote computer-storage media including memory-storage devices.
  • FIG. 2 An illustrative diagram of a computer program product (i.e., storage device) embodying the present invention is depicted in FIG. 2.
  • the computer program product is depicted as floppy disk 200 or an optical disk 202 such as a CD or DVD.
  • the computer program product generally represents computer-readable instructions stored on any compatible non-transitory computer-readable medium.
  • the term "instructions” as used with respect to this invention generally indicates a set of operations to be performed on a computer, and may represent pieces of a whole program or individual, separable, software modules.
  • Non-limiting examples of "instruction” include computer program code (source or object code) and "hard-coded" electronics (i.e. computer operations coded into a computer chip).
  • the "instruction" is stored on any non-transitory computer-readable medium, such as in the memory of a computer or on a floppy disk, a CD-ROM, and a flash drive. In either event, the instructions are encoded on a non-transitory computer-readable medium.
  • the system comprises an attribute aware joint dictionary learning mechanism with a unique attribute- aware formulation for the zero-shot learning (ZSL) problem, which significantly improves the zero-shot performance of the system over existing techniques.
  • ZSL zero-shot learning
  • the system contains a training and a testing phase.
  • the training phase the system takes as input a set of training images containing objects from many classes of interest, where each object image has been annotated with both class labels (e.g., "bear”, “car”, “house”) and several semantic attributes (e.g., "can fly”, “has wheels", “green”).
  • Class labels are typically the nouns or words that would be used to define or describe the object.
  • Semantic attributes are a verbal definion of the object (often an adjective, phrase, or sentence) that contributes to the meaning of the object.
  • attributes to train a model that maps low level image features (such as edges, corners, and gradients) to semantic attributes.
  • the mapping is learned in a way to capture the shared information between image features and attributes in an optimal manner.
  • the system takes as input an image of an object not in the training set (i.e., no instance of the object is included in the training phase), and uses the learned model to map test image features to semantic attributes, so that, given a semantic description for the test object, the test image can be correctly classified despite the test object not having been present in the training image set.
  • the system includes: 1) modeling the relationship between visual features and semantic attributes using joint sparse representations with respect to dictionaries of visual features and semantic attributes; and 2) an entropy regularization for joint dictionaries that significantly increases the fidelity of the learned representations and improves performance of existing technologies (i.e., an improvement over the state-of-the-art) on publicly available datasets.
  • the system described herein minimizes the need for labeled data in supervised learning via knowledge transfer by finding a mapping from the visual data to a semantic attribute space.
  • the zero-shot learning (ZSL) paradigm aims at classifying previously unseen data classes.
  • the need for ZSL arises mainly from a lack of annotated data, together with the constant emergence of new visual categories (e.g., new products, new models of vehicles, etc.).
  • the system described herein provides a ZSL machine-vision system that incorporates an attribute-aware joint sparse dictionary learning to model the relationship between visual features of an object and its semantic attributes.
  • the assumption behind ZSL methods is that the training (i.e. seen) and testing (i.e.
  • semantic attributes are often provided as accessible side information (e.g., a word description of the classes), which uniquely describes a data class.
  • the training phase the relationship between the seen data and its corresponding attributes is learned. Consequently, in the testing phase, the input data from an unseen class is parsed into its attributes and the label is predicted from these extracted attributes.
  • the invention described herein improves the zero-shot capability of a machine vision system compared to the state-of-the-art by leveraging a mathematically rigorous model that encodes the relationship between an object and its semantic attributes.
  • a purpose of the invention is to recognize novel objects or scenes in camera images.
  • the camera may be electric optical, infrared (IR), short wave IR, or similar, such that the output is an intensity image with one or more color-like channels.
  • IR infrared
  • short wave IR short wave IR
  • the attribute-aware joint dictionary learning method according to embodiments of the present disclosure is a natural fit for this purpose.
  • the feature space and the attribute space are modeled as nonlinear spaces characterized by a union of low-dimensional (i.e., dimensions significantly smaller than the dimension of the feature space) linear spaces.
  • the two nonlinear spaces are constrained to have homologous components, hence the name joint dictionaries, and they are modeled to provide the same representation for image features and their corresponding attributes.
  • the training phase of the zero-shot machine vision system is depicted in FIG. 3.
  • the image features (element 300) are extracted from a deep convolutional neural network, while word2vec (see Literature Reference No. 4) is used to extract attributes (element 302) from the verbal description of the class.
  • the joint dictionary learning approach (element 304) enforces that the sparse representations of features (element 306) and of their corresponding attributes (element 308) be the same sparse representations.
  • CNN Convolutional Neural Network
  • N is the number of images, and each image has an associated P- dimensional feature vector.
  • Z [z 1 ; . . . , Z N ] 6 R® xN represent the set of corresponding attributes (element 302) for the images, each element of Z being a Q-dimensional attribute vector.
  • the word ' attributes' is used in its broadest sense, which encompasses word embeddings or any other semantic information for the images.
  • Z' [z ⁇ , . . . , z M ' ] E R Q xM be attributes of previously unseen classes of images, where M is the number of such classes.
  • the label for the i'th image is denoted as yi E R , where the camera image can potentially have multiple memberships of the K classes.
  • the semantic attributes contain information of all seen (Z) and unseen ( ⁇ ') images, including the zebras. Semantic attributes in this case could be the verbal definition of animals (e.g., ' a zebra is a white horse-like animal with black tiger-like stripes'). It can be seen that, by learning the relationship between the image features and the attributes ' horse-like' and ' has stripes' from the seen images, one should be able to assign the previously unseen zebra image to its corresponding attribute.
  • D G R PX L is the image feature dictionary
  • a G R L X N is the joint
  • Equations (1) and (2) are not convex in D X , A) and ⁇ D Z , 5) respectively; however, they are convex in each individual parameter given the rest.
  • an iterative scheme was devised to solve for one of the dictionaries (i.e., D X and Z ) z ) at a time and fixing the other, until convergence is achieved. Convergence to local optima is guaranteed.
  • Each optimization is minimized by solving a Lasso problem (see Literature Reference No. 6 for a description of the Lasso problem) to find the sparse representations followed by a Quadratically Constrained Quadratic Program (QCQP) (see Literature Reference No. 7) to update the dictionaries. The steps are then repeated.
  • a Lasso problem see Literature Reference No. 6 for a description of the Lasso problem
  • QQP Quadratically Constrained Quadratic Program
  • X G R is first represented as linear combinations of the atoms of dictionary
  • FIG. 4 shows a schematic of the testing phase for an input image (element 400), the image features (element 402) are first extracted and a Lasso problem is solved to find the sparse representation of the image (element 404). Given that the sparse representation is shared among features and attributes, image attributes are estimated (element 406), and the closest attribute in the attribute space (element 408) is found.
  • Z the probability of prediction
  • Z the probability of prediction
  • z e.g., ⁇ (element 502) or z2' (element 504)
  • z * argmin zE[z z > ⁇ ⁇ z - D z a * ⁇ 2 2 .
  • the system addresses the need for robust machine vision systems on autonomous platforms (e.g., drones, unmanned aerial vehicles (UAVs)) and autonomous vehicles.
  • autonomous platforms e.g., drones, unmanned aerial vehicles (UAVs)
  • UAVs unmanned aerial vehicles
  • the invention described herein provides the capability of effectively adapting to novel scenarios (e.g., novel objects, weather conditions, etc.) and enables future transfer learning technologies.
  • the system can generate an alert when an avoidance object is detected, such that the autonomous platform is caused to perform an automatic operation, such as a braking or swerving operation to avoid hitting the object.
  • the alert can be an audible alert (e.g., beep, tone) and/or a visual alert (e.g., light or message on dashboard).
  • the system may cause the autonomous vehicle to apply a functional response, such as a braking operation, to stop the vehicle.
  • a functional response such as a braking operation
  • Other appropriate responses may include one or more of a steering operation, a throttle operation to increase speed or to decrease speed, or a decision to maintain course and speed without change.
  • the responses may be appropriate for avoiding a collision, improving travel speed, or improving efficiency.
  • FIG. 7 is a flow diagram illustrating using a processor 700 to control a
  • classifications of unseen images include a vehicle or a vehicle component, such as a brake, a steering mechanism, suspension, or safety device (e.g., airbags, seatbelt tensioners, etc.).
  • vehicle could be an unmanned aerial vehicle (UAV), an autonomous ground vehicle, or a human operated vehicle controlled either by a driver or by a remote operator.
  • UAV unmanned aerial vehicle
  • autonomous ground vehicle or a human operated vehicle controlled either by a driver or by a remote operator.
  • control of other device types is also possible.
  • the attribute aware joint sparse visual feature and sematic attribute modeling system of this disclosure enables one to perform zero-shot machine vision with far fewer training examples than existing systems.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computational Linguistics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

La présente invention concerne un système de reconnaissance d'objet. Le système génère un ensemble d'images d'apprentissage d'images d'objets à partir de multiples catégories d'images. Un modèle apprend à l'aide d'un ensemble d'images d'apprentissage et d'attributs sémantiques annotés, le modèle mettant en correspondance des caractéristiques visuelles provenant d'images connues avec les attributs sémantiques annotés à l'aide de représentations éparses communes par rapport à des dictionnaires de caractéristiques visuelles et d'attributs sémantiques. Le modèle ayant appris sert à mettre en correspondance des caractéristiques visuelles d'une image d'entrée jamais vue avec des attributs sémantiques. L'image d'entrée jamais vue est classée comme appartenant à une catégorie d'images, et un dispositif est commandé sur la base de la classification de l'image d'entrée jamais vue.
PCT/US2018/041806 2017-09-12 2018-07-12 Système visionique sans prise de vue sensible aux attributs par l'intermédiaire de représentations éparses communes WO2019055114A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP18856563.4A EP3682370A4 (fr) 2017-09-12 2018-07-12 Système visionique sans prise de vue sensible aux attributs par l'intermédiaire de représentations éparses communes
CN201880052204.6A CN111052144A (zh) 2017-09-12 2018-07-12 借由联合稀疏表示的属性感知零样本机器视觉系统

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US201762557721P 2017-09-12 2017-09-12
US62/557,721 2017-09-12
US15/949,896 2018-04-10
US15/949,896 US10755149B2 (en) 2017-05-05 2018-04-10 Zero shot machine vision system via joint sparse representations

Publications (1)

Publication Number Publication Date
WO2019055114A1 true WO2019055114A1 (fr) 2019-03-21

Family

ID=65723016

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2018/041806 WO2019055114A1 (fr) 2017-09-12 2018-07-12 Système visionique sans prise de vue sensible aux attributs par l'intermédiaire de représentations éparses communes

Country Status (3)

Country Link
EP (1) EP3682370A4 (fr)
CN (1) CN111052144A (fr)
WO (1) WO2019055114A1 (fr)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110135459A (zh) * 2019-04-15 2019-08-16 天津大学 一种基于双三元组深度度量学习网络的零样本分类方法
CN110580501A (zh) * 2019-08-20 2019-12-17 天津大学 一种基于变分自编码对抗网络的零样本图像分类方法
CN110807472A (zh) * 2019-10-12 2020-02-18 北京达佳互联信息技术有限公司 图像识别方法、装置、电子设备及存储介质
CN111325347A (zh) * 2020-02-19 2020-06-23 山东大学 基于可解释视觉推理模型的危险预警描述自动生成方法
CN112364893A (zh) * 2020-10-23 2021-02-12 天津大学 一种基于数据增强的半监督零样本图像分类方法
CN112364894A (zh) * 2020-10-23 2021-02-12 天津大学 一种基于元学习的对抗网络的零样本图像分类方法
CN112418257A (zh) * 2019-08-22 2021-02-26 四川大学 一种有效的基于潜在视觉属性挖掘的零样本学习方法
CN113688879A (zh) * 2021-07-30 2021-11-23 南京理工大学 一种基于置信度分布外检测的广义零样本学习分类方法
CN115424096A (zh) * 2022-11-08 2022-12-02 南京信息工程大学 一种多视角零样本图像识别方法
CN116109877A (zh) * 2023-04-07 2023-05-12 中国科学技术大学 组合式零样本图像分类方法、系统、设备及存储介质
WO2023246641A1 (fr) * 2022-06-24 2023-12-28 华为云计算技术有限公司 Procédé et appareil d'identification d'objet, et support d'enregistrement
CN113688879B (zh) * 2021-07-30 2024-05-24 南京理工大学 一种基于置信度分布外检测的广义零样本学习分类方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112163603A (zh) * 2020-09-15 2021-01-01 郑州金惠计算机系统工程有限公司 零样本图像识别方法、装置、电子设备及存储介质
CN114925802A (zh) * 2021-12-27 2022-08-19 天翼数字生活科技有限公司 一种基于深度特征映射的集成迁移学习方法和系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104036021A (zh) * 2014-06-26 2014-09-10 广西师范大学 混合生成式和判别式学习模型的图像语义标注方法
US9110922B2 (en) * 2010-02-01 2015-08-18 Google Inc. Joint embedding for item association
US20160239711A1 (en) * 2013-10-18 2016-08-18 Vision Semanatics Limited Visual Data Mining
US9566981B2 (en) * 2014-09-01 2017-02-14 Honda Research Institute Europe Gmbh Method and system for post-collision manoeuvre planning and vehicle equipped with such system
WO2017139764A1 (fr) * 2016-02-12 2017-08-17 Sri International Détection d'événement sans exemple d'apprentissage par intégration sémantique

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030202683A1 (en) * 2002-04-30 2003-10-30 Yue Ma Vehicle navigation system that automatically translates roadside signs and objects
CN105389584B (zh) * 2015-10-13 2018-07-10 西北工业大学 基于卷积神经网络与语义转移联合模型的街景语义标注方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9110922B2 (en) * 2010-02-01 2015-08-18 Google Inc. Joint embedding for item association
US20160239711A1 (en) * 2013-10-18 2016-08-18 Vision Semanatics Limited Visual Data Mining
CN104036021A (zh) * 2014-06-26 2014-09-10 广西师范大学 混合生成式和判别式学习模型的图像语义标注方法
US9566981B2 (en) * 2014-09-01 2017-02-14 Honda Research Institute Europe Gmbh Method and system for post-collision manoeuvre planning and vehicle equipped with such system
WO2017139764A1 (fr) * 2016-02-12 2017-08-17 Sri International Détection d'événement sans exemple d'apprentissage par intégration sémantique

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3682370A4 *

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110135459A (zh) * 2019-04-15 2019-08-16 天津大学 一种基于双三元组深度度量学习网络的零样本分类方法
CN110135459B (zh) * 2019-04-15 2023-04-07 天津大学 一种基于双三元组深度度量学习网络的零样本分类方法
CN110580501A (zh) * 2019-08-20 2019-12-17 天津大学 一种基于变分自编码对抗网络的零样本图像分类方法
CN110580501B (zh) * 2019-08-20 2023-04-25 天津大学 一种基于变分自编码对抗网络的零样本图像分类方法
CN112418257A (zh) * 2019-08-22 2021-02-26 四川大学 一种有效的基于潜在视觉属性挖掘的零样本学习方法
CN112418257B (zh) * 2019-08-22 2023-04-18 四川大学 一种有效的基于潜在视觉属性挖掘的零样本学习方法
CN110807472B (zh) * 2019-10-12 2022-08-12 北京达佳互联信息技术有限公司 图像识别方法、装置、电子设备及存储介质
CN110807472A (zh) * 2019-10-12 2020-02-18 北京达佳互联信息技术有限公司 图像识别方法、装置、电子设备及存储介质
CN111325347A (zh) * 2020-02-19 2020-06-23 山东大学 基于可解释视觉推理模型的危险预警描述自动生成方法
CN112364893A (zh) * 2020-10-23 2021-02-12 天津大学 一种基于数据增强的半监督零样本图像分类方法
CN112364894B (zh) * 2020-10-23 2022-07-08 天津大学 一种基于元学习的对抗网络的零样本图像分类方法
CN112364894A (zh) * 2020-10-23 2021-02-12 天津大学 一种基于元学习的对抗网络的零样本图像分类方法
CN113688879A (zh) * 2021-07-30 2021-11-23 南京理工大学 一种基于置信度分布外检测的广义零样本学习分类方法
CN113688879B (zh) * 2021-07-30 2024-05-24 南京理工大学 一种基于置信度分布外检测的广义零样本学习分类方法
WO2023246641A1 (fr) * 2022-06-24 2023-12-28 华为云计算技术有限公司 Procédé et appareil d'identification d'objet, et support d'enregistrement
CN115424096A (zh) * 2022-11-08 2022-12-02 南京信息工程大学 一种多视角零样本图像识别方法
CN115424096B (zh) * 2022-11-08 2023-01-31 南京信息工程大学 一种多视角零样本图像识别方法
CN116109877A (zh) * 2023-04-07 2023-05-12 中国科学技术大学 组合式零样本图像分类方法、系统、设备及存储介质
CN116109877B (zh) * 2023-04-07 2023-06-20 中国科学技术大学 组合式零样本图像分类方法、系统、设备及存储介质

Also Published As

Publication number Publication date
EP3682370A4 (fr) 2021-06-09
EP3682370A1 (fr) 2020-07-22
CN111052144A (zh) 2020-04-21

Similar Documents

Publication Publication Date Title
US10908616B2 (en) Attribute aware zero shot machine vision system via joint sparse representations
WO2019055114A1 (fr) Système visionique sans prise de vue sensible aux attributs par l'intermédiaire de représentations éparses communes
US10755149B2 (en) Zero shot machine vision system via joint sparse representations
CN110073367B (zh) 用于零样本学习的利用基于soft-max的兼容性函数的多视图嵌入
CN108604303B (zh) 用于场景分类的系统、方法和计算机可读介质
US11448753B2 (en) System and method for transferring electro-optical (EO) knowledge for synthetic-aperture-radar (SAR)-based object detection
US9922265B2 (en) Global-scale object detection using satellite imagery
US9031331B2 (en) Metric learning for nearest class mean classifiers
CN109997152B (zh) 利用多尺度流形对准的零样本学习
Yakhnenko et al. Annotating images and image objects using a hierarchical dirichlet process model
Kolouri et al. Joint dictionaries for zero-shot learning
CN111492375B (zh) 用于辨识新对象的机器视觉系统
WO2021011062A1 (fr) Système et procédé d'apprentissage continu à l'aide de relecture d'expérience
Lu et al. Beyond bag of latent topics: spatial pyramid matching for scene category recognition
CN116997908A (zh) 用于分类类型任务的连续学习神经网络系统训练
US20220230425A1 (en) Object discovery in images through categorizing object parts
WO2021096797A1 (fr) Apprentissage de représentation de caractéristique universelle pour reconnaissance faciale
CN113780331A (zh) 由计算机实施的训练方法、分类方法及系统和计算机可读记录介质
US20210383226A1 (en) Cross-transformer neural network system for few-shot similarity determination and classification
CN113316790B (zh) 用于自主学习代理中的无监督域适应的系统、方法和介质
Li et al. Weakly-supervised scene parsing with multiple contextual cues
Tian et al. Robust manifold learning based ordinal discriminative correlation regression
Flenner et al. Learning representations for improved target identification, scene classification, and information fusion
CN116524288A (zh) 用于训练模型的方法、设备和程序产品
Casterline et al. Semantically enabled image similarity search

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18856563

Country of ref document: EP

Kind code of ref document: A1

DPE1 Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101)
NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2018856563

Country of ref document: EP

Effective date: 20200414