CN114882441A - Pedestrian re-identification method and system based on deep learning - Google Patents

Pedestrian re-identification method and system based on deep learning Download PDF

Info

Publication number
CN114882441A
CN114882441A CN202210599800.4A CN202210599800A CN114882441A CN 114882441 A CN114882441 A CN 114882441A CN 202210599800 A CN202210599800 A CN 202210599800A CN 114882441 A CN114882441 A CN 114882441A
Authority
CN
China
Prior art keywords
pedestrian
recognition
training
identification
recognition model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210599800.4A
Other languages
Chinese (zh)
Inventor
吕雪
陈雪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhengzhou University
Original Assignee
Zhengzhou University
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
Application filed by Zhengzhou University filed Critical Zhengzhou University
Priority to CN202210599800.4A priority Critical patent/CN114882441A/en
Publication of CN114882441A publication Critical patent/CN114882441A/en
Pending legal-status Critical Current

Links

Images

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
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/761—Proximity, similarity or dissimilarity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Databases & Information Systems (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Mathematical Physics (AREA)
  • Medical Informatics (AREA)
  • Biomedical Technology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a pedestrian re-identification method and system based on deep learning. The method comprises the following steps: collecting a monitoring video, and decomposing an image frame from the monitoring video; acquiring an image frame set which contains a target pedestrian and is clear without shielding from the image frame, and performing feature extraction on the image frame set by using a convolutional neural network of a pedestrian re-identification model to obtain the features of the target pedestrian; training the pedestrian re-recognition model according to the target pedestrian characteristics, determining the training loss of the pedestrian re-recognition model, optimizing the parameters of the pedestrian re-recognition model according to the training loss, and performing pedestrian re-recognition by using the optimized pedestrian re-recognition model. By adopting the technical scheme, the pedestrian re-identification can be accurately carried out, and the accuracy of pedestrian identification is improved.

Description

Pedestrian re-identification method and system based on deep learning
Technical Field
The invention relates to the technical field of neural networks, in particular to a pedestrian re-identification method and system based on deep learning.
Background
Pedestrian re-identification (Person re-identification), also known as pedestrian re-identification, is a technique that uses computer vision techniques to determine whether a particular pedestrian is present in an image or video sequence. Is widely considered as a sub-problem for image retrieval. Given a monitored pedestrian image, the pedestrian image is retrieved across the device. The visual limitation of a fixed camera is overcome, the pedestrian detection/pedestrian tracking technology can be combined, and the method can be widely applied to the fields of intelligent video monitoring, intelligent security and the like.
However, the existing pedestrian re-identification method only processes the acquired video image and then performs operations such as image positioning and the like, and the pedestrian re-identification method has low efficiency and low accuracy. Based on the method, the invention provides a pedestrian re-identification method and system based on deep learning.
Disclosure of Invention
The invention provides a pedestrian re-identification method based on deep learning, which comprises the following steps:
collecting a monitoring video, and decomposing an image frame from the monitoring video;
acquiring an image frame set which contains a target pedestrian and is clear without shielding from the image frame, and performing feature extraction on the image frame set by using a convolutional neural network of a pedestrian re-identification model to obtain the features of the target pedestrian;
training the pedestrian re-recognition model according to the target pedestrian characteristics, determining the training loss of the pedestrian re-recognition model, optimizing the parameters of the pedestrian re-recognition model according to the training loss, and performing pedestrian re-recognition by using the optimized pedestrian re-recognition model.
In the method for re-identifying pedestrians based on deep learning, image frames decomposed from a surveillance video are used as training image frames, the training image sets are part of the acquired image sets which are randomly divided, and the selection of the training image sets is diversified, and may include pedestrian posture diversity, pedestrian scale diversity and background diversity, for example.
The pedestrian re-identification method based on deep learning is characterized in that pedestrian detection is realized in a manual marking or automatic marking mode, and a training image set is preprocessed to obtain an image frame set which contains a target pedestrian and is clear without occlusion.
The pedestrian re-recognition method based on deep learning, wherein the training of the pedestrian re-recognition model according to the target pedestrian features specifically comprises:
processing the original features by using an attention module of a pedestrian re-identification model to obtain a plurality of pedestrian local features;
determining a similarity matrix between local features of each pedestrian by using a neural network of a pedestrian re-identification model, and adjusting the local features of each pedestrian according to the similarity matrix;
determining a pedestrian recognition result based on the adjusted local features of the pedestrian, determining the training loss of the pedestrian re-recognition model according to the pedestrian recognition result, and optimizing the parameters of the pedestrian re-recognition model according to the training loss.
After the adjusted local features of the pedestrian are obtained, classifying the pedestrian IDs in the image by using a classification network in a pedestrian re-identification model based on the local features of the pedestrian, and further obtaining a pedestrian re-identification result and a corresponding classification loss. The classification network can adopt a softmax classifier, the pedestrian ID classification loss of the softmax classifier is correspondingly obtained, based on the classification loss, the parameters of the pedestrian re-identification model are optimized, and finally the trained pedestrian re-identification model is obtained.
The pedestrian re-recognition method based on deep learning specifically comprises the following substeps of optimizing parameters of the pedestrian re-recognition model according to training loss:
inputting the pedestrian re-recognition training data set into a preset pedestrian re-recognition model for processing, and calculating a first loss function based on a processing result;
performing enhancement processing on the first pedestrian re-recognition training data set, inputting a processing result into a supervision network for processing, and calculating a second loss function;
and adjusting the pedestrian re-recognition model according to the calculation results of the first loss function and the second loss function to obtain the target pedestrian re-recognition model.
The invention provides a pedestrian re-identification system based on deep learning, which comprises:
the image acquisition module is used for acquiring a monitoring video and decomposing an image frame from the monitoring video;
the target pedestrian feature extraction module is used for acquiring an image frame set which contains a target pedestrian and is clear without shielding from image frames, and performing feature extraction on the image frame set by utilizing a convolutional neural network of a pedestrian re-identification model to obtain target pedestrian features;
and the pedestrian re-recognition module is used for training the pedestrian re-recognition model according to the target pedestrian characteristics, determining the training loss of the pedestrian re-recognition model, optimizing the parameters of the pedestrian re-recognition model according to the training loss, and performing pedestrian re-recognition by using the optimized pedestrian re-recognition model.
The pedestrian re-identification system based on deep learning is characterized in that the pedestrian re-identification module is specifically configured to process the original features by using an attention module of a pedestrian re-identification model to obtain a plurality of local features of pedestrians; determining a similarity matrix between local features of each pedestrian by using a neural network of a pedestrian re-identification model, and adjusting the local features of each pedestrian according to the similarity matrix; determining a pedestrian recognition result based on the adjusted local features of the pedestrian, determining the training loss of the pedestrian re-recognition model according to the pedestrian recognition result, and optimizing the parameters of the pedestrian re-recognition model according to the training loss.
The pedestrian re-recognition system based on deep learning is characterized in that the pedestrian re-recognition module is specifically configured to input a pedestrian re-recognition training data set into a preset pedestrian re-recognition model for processing, and calculate a first loss function based on a processing result; performing enhancement processing on the first pedestrian re-recognition training data set, inputting a processing result into a supervision network for processing, and calculating a second loss function; and adjusting the pedestrian re-recognition model according to the calculation results of the first loss function and the second loss function to obtain the target pedestrian re-recognition model.
The invention has the following beneficial effects: by adopting the technical scheme, the pedestrian re-identification can be accurately carried out, and the accuracy of pedestrian identification is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings can be obtained by those skilled in the art according to the drawings.
Fig. 1 is a flowchart of a pedestrian re-identification method based on deep learning according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
As shown in fig. 1, an embodiment of the present invention provides a pedestrian re-identification method based on deep learning, including:
step 110, collecting a monitoring video, and decomposing an image frame from the monitoring video;
specifically, a surveillance video is collected from a certain road segment camera device. The image frames decomposed from the surveillance video are used as training image frames, and these training image sets may be randomly divided from the acquired image sets, for example, 50% of the image sets are randomly divided as the training image sets. In order to improve the accuracy of pedestrian re-recognition, the selection of the training image set should be various, and may include, for example, pedestrian posture diversity, pedestrian scale diversity, background diversity, and the like.
Step 120, obtaining an image frame set which contains a target pedestrian and is clear without shielding from the image frame, and performing feature extraction on the image frame set by using a convolutional neural network of a pedestrian re-identification model to obtain the features of the target pedestrian;
in the embodiment of the application, pedestrian detection can be realized in a manual marking or automatic marking mode, and the training image set is preprocessed to obtain the clear image frame set which contains target pedestrians and is free of shielding. When the pedestrian re-identification model is trained, a certain number of pedestrian images can be obtained first, then the obtained pedestrian images are subjected to convolution processing by using a convolution network, and original features corresponding to the pedestrian images are extracted. The convolution network can adopt a residual error network ResNet, wherein ResNet is mainly used for solving the problem of network degradation caused by the increase of the network depth in the training of the deep neural network, and the method is mainly used for extracting the target feature in the embodiment of the application.
And step 130, training the pedestrian re-recognition model according to the target pedestrian characteristics, determining the training loss of the pedestrian re-recognition model, optimizing the parameters of the pedestrian re-recognition model according to the training loss, and performing pedestrian re-recognition by using the optimized pedestrian re-recognition model.
Specifically, training the pedestrian re-recognition model according to the target pedestrian features specifically comprises:
step 131, processing the original features by using an attention module of a pedestrian re-identification model to obtain a plurality of pedestrian local features;
in a pedestrian re-recognition scene, a pedestrian region in an image is an image region which we want to pay more attention to, so the embodiment of the application introduces an attention (attention) module as a basis for training a pedestrian re-recognition model, and an attention mechanism simulates an internal process of biological observation behavior and is a mechanism for aligning internal experience and external feeling so as to increase the observation fineness of a partial region. The attention mechanism can quickly extract important features of sparse data, and therefore, the attention mechanism is widely used in the fields of natural language processing tasks and image processing tasks. After the original features of the pedestrian image are obtained, the original features are further processed by an attention module, and features which are more beneficial to pedestrian re-identification and more obvious in the original features are extracted to serve as local features of the pedestrian.
And 132, determining a similarity matrix between the local features of each pedestrian by using a neural network of the pedestrian re-identification model, and adjusting the local features of each pedestrian according to the similarity matrix.
Step 133, determining a pedestrian recognition result based on the adjusted local features of the pedestrian, determining a training loss of the pedestrian re-recognition model according to the pedestrian recognition result, and optimizing parameters of the pedestrian re-recognition model according to the training loss;
after the adjusted pedestrian local features are obtained, classifying the pedestrian IDs in the image by utilizing a classification network in a pedestrian re-identification model based on the pedestrian local features, and further obtaining a pedestrian re-identification result and corresponding classification loss. The classification network can adopt a softmax classifier, the pedestrian ID classification loss of the softmax classifier is correspondingly obtained, based on the classification loss, the parameters of the pedestrian re-identification model are optimized, and finally the trained pedestrian re-identification model is obtained.
The method specifically comprises the following substeps of optimizing parameters of the pedestrian re-identification model according to training loss:
step 133-1, inputting the pedestrian re-recognition training data set into a preset pedestrian re-recognition model for processing, and calculating a first loss function based on a processing result;
the preset pedestrian re-recognition model can adopt a currently known pedestrian re-recognition network structure or a network structure developed in the future, and can be trained by applying the training method. The first loss function may use various loss functions suitable for training of the pedestrian re-recognition model, such as a contrast loss, a triplet loss, a quadruplet loss, and the like.
And 133-2, performing enhancement processing on the first pedestrian re-recognition training data set, inputting a processing result into a supervision network for processing, and calculating a second loss function.
The enhancement processing includes contrast and brightness enhancement processing. The enhancement process may also be a variety of geometric transformation processes. The supervision network is used for extracting supervision information from the pedestrian re-recognition training data set so as to assist training of the pedestrian re-recognition model. The supervised information may be reflected in a loss function of the supervised network, thus assisting the training of the pedestrian re-identification model by the second loss function, and forming a positive contribution to the training result resultant with different training data sets due to the additional common supervised information provided by the supervised network.
And step 133-3, adjusting the pedestrian re-recognition model according to the calculation results of the first loss function and the second loss function to obtain the target pedestrian re-recognition model.
After the first loss function and the second loss function are calculated, adjusting a preset pedestrian re-identification model according to the calculation results of the first loss function and the second loss function, wherein the adjustment includes adjusting network parameters and/or weights of the preset pedestrian re-identification model, and the like, until the calculation result of the loss function reaches a set threshold value or the calculation result of the loss function is not reduced any more, so that the target pedestrian re-identification model is obtained.
Example two
The embodiment of the invention provides a pedestrian re-identification system based on deep learning, which comprises:
the image acquisition module is used for acquiring a monitoring video and decomposing an image frame from the monitoring video;
the target pedestrian feature extraction module is used for acquiring an image frame set which contains a target pedestrian and is clear without shielding from image frames, and performing feature extraction on the image frame set by utilizing a convolutional neural network of a pedestrian re-identification model to obtain target pedestrian features;
and the pedestrian re-recognition module is used for training the pedestrian re-recognition model according to the target pedestrian characteristics, determining the training loss of the pedestrian re-recognition model, optimizing the parameters of the pedestrian re-recognition model according to the training loss, and performing pedestrian re-recognition by using the optimized pedestrian re-recognition model.
The pedestrian re-identification module is specifically used for processing the original features by utilizing an attention module of the pedestrian re-identification module to obtain a plurality of pedestrian local features; determining a similarity matrix between local features of each pedestrian by using a neural network of a pedestrian re-identification model, and adjusting the local features of each pedestrian according to the similarity matrix; determining a pedestrian recognition result based on the adjusted local features of the pedestrian, determining the training loss of the pedestrian re-recognition model according to the pedestrian recognition result, and optimizing the parameters of the pedestrian re-recognition model according to the training loss.
The pedestrian re-recognition module is specifically used for inputting the pedestrian re-recognition training data set into a preset pedestrian re-recognition model for processing, and calculating a first loss function based on the processing result; performing enhancement processing on the first pedestrian re-recognition training data set, inputting a processing result into a supervision network for processing, and calculating a second loss function; and adjusting the pedestrian re-recognition model according to the calculation results of the first loss function and the second loss function to obtain the target pedestrian re-recognition model.
In accordance with the embodiments described above, embodiments of the present invention provide a computer-readable storage medium having one or more program instructions embodied therein, the one or more program instructions being for execution by a processor to perform a method for pedestrian re-identification based on deep learning.
The disclosed embodiments of the present invention provide a computer-readable storage medium having computer program instructions stored therein, which when run on a computer, cause the computer to perform a pedestrian re-identification method based on deep learning as described above.
In an embodiment of the invention, the processor may be an integrated circuit chip having signal processing capability. The Processor may be a general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete Gate or transistor logic device, discrete hardware component.
The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The processor reads the information in the storage medium and completes the steps of the method in combination with the hardware.
The storage medium may be a memory, for example, which may be volatile memory or nonvolatile memory, or which may include both volatile and nonvolatile memory.
The nonvolatile Memory may be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable PROM (EEPROM), or a flash Memory.
The volatile Memory may be a Random Access Memory (RAM) which serves as an external cache. By way of example, and not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), SLDRAM (SLDRAM), and Direct Rambus RAM (DRRAM).
The storage media described in connection with the embodiments of the invention are intended to comprise, without being limited to, these and any other suitable types of memory.
Those skilled in the art will appreciate that the functionality described in the present invention may be implemented in a combination of hardware and software in one or more of the examples described above. When software is applied, the corresponding functionality may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a general purpose or special purpose computer.
The above-mentioned embodiments, objects, technical solutions and advantages of the present invention are further described in detail, it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made on the basis of the technical solutions of the present invention should be included in the scope of the present invention.

Claims (9)

1. A pedestrian re-identification method based on deep learning is characterized by comprising the following steps:
collecting a monitoring video, and decomposing an image frame from the monitoring video;
acquiring an image frame set which contains a target pedestrian and is clear without shielding from the image frame, and performing feature extraction on the image frame set by using a convolutional neural network of a pedestrian re-identification model to obtain the features of the target pedestrian;
training the pedestrian re-recognition model according to the target pedestrian characteristics, determining the training loss of the pedestrian re-recognition model, optimizing the parameters of the pedestrian re-recognition model according to the training loss, and performing pedestrian re-recognition by using the optimized pedestrian re-recognition model.
2. The method as claimed in claim 1, wherein the image frames separated from the surveillance video are training image frames, the training image sets are randomly separated from the acquired image sets, and the selection of the training image sets is diverse, such as pedestrian pose diversity, pedestrian scale diversity and background diversity.
3. The deep learning-based pedestrian re-identification method as claimed in claim 1, wherein the pedestrian detection is realized by manual labeling or automatic labeling, and the training image set is preprocessed to obtain the image frame set which contains the target pedestrian and is clear without occlusion.
4. The pedestrian re-recognition method based on deep learning as claimed in claim 1, wherein training the pedestrian re-recognition model according to the target pedestrian features specifically comprises:
processing the original features by using an attention module of a pedestrian re-identification model to obtain a plurality of pedestrian local features;
determining a similarity matrix between local features of each pedestrian by using a neural network of a pedestrian re-identification model, and adjusting the local features of each pedestrian according to the similarity matrix;
determining a pedestrian recognition result based on the adjusted local features of the pedestrian, determining the training loss of the pedestrian re-recognition model according to the pedestrian recognition result, and optimizing the parameters of the pedestrian re-recognition model according to the training loss.
5. The deep learning-based pedestrian re-identification method according to claim 1, wherein after the adjusted local features of the pedestrian are obtained, the pedestrian ID in the image is classified by using a classification network in a pedestrian re-identification model based on the local features of the pedestrian, so as to obtain a result of the pedestrian re-identification and a corresponding classification loss, wherein the classification network can adopt a softmax classifier to obtain a pedestrian ID classification loss of the softmax classifier correspondingly, and based on the classification loss, parameters of the pedestrian re-identification model are optimized, so as to finally obtain the trained pedestrian re-identification model.
6. The pedestrian re-recognition method based on deep learning as claimed in claim 1, wherein the optimization of the parameters of the pedestrian re-recognition model according to the training loss specifically comprises the following sub-steps:
inputting the pedestrian re-recognition training data set into a preset pedestrian re-recognition model for processing, and calculating a first loss function based on a processing result;
performing enhancement processing on the first pedestrian re-recognition training data set, inputting a processing result into a supervision network for processing, and calculating a second loss function;
and adjusting the pedestrian re-recognition model according to the calculation results of the first loss function and the second loss function to obtain the target pedestrian re-recognition model.
7. A pedestrian re-identification system based on deep learning is characterized by comprising:
the image acquisition module is used for acquiring a monitoring video and decomposing an image frame from the monitoring video;
the target pedestrian feature extraction module is used for acquiring an image frame set which contains a target pedestrian and is clear without shielding from image frames, and performing feature extraction on the image frame set by utilizing a convolutional neural network of a pedestrian re-identification model to obtain target pedestrian features;
and the pedestrian re-recognition module is used for training the pedestrian re-recognition model according to the target pedestrian characteristics, determining the training loss of the pedestrian re-recognition model, optimizing the parameters of the pedestrian re-recognition model according to the training loss, and performing pedestrian re-recognition by using the optimized pedestrian re-recognition model.
8. The deep learning-based pedestrian re-identification system according to claim 7, wherein the pedestrian re-identification module is specifically configured to process the original features by using an attention module of a pedestrian re-identification model to obtain a plurality of pedestrian local features; determining a similarity matrix between local features of each pedestrian by using a neural network of a pedestrian re-identification model, and adjusting the local features of each pedestrian according to the similarity matrix; determining a pedestrian recognition result based on the adjusted local features of the pedestrian, determining the training loss of the pedestrian re-recognition model according to the pedestrian recognition result, and optimizing the parameters of the pedestrian re-recognition model according to the training loss.
9. The deep learning-based pedestrian re-identification system according to claim 7, wherein the pedestrian re-identification module is specifically configured to input the pedestrian re-identification training data set into a preset pedestrian re-identification model for processing, and calculate the first loss function based on the processing result; performing enhancement processing on the first pedestrian re-recognition training data set, inputting a processing result into a supervision network for processing, and calculating a second loss function; and adjusting the pedestrian re-recognition model according to the calculation results of the first loss function and the second loss function to obtain the target pedestrian re-recognition model.
CN202210599800.4A 2022-05-30 2022-05-30 Pedestrian re-identification method and system based on deep learning Pending CN114882441A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210599800.4A CN114882441A (en) 2022-05-30 2022-05-30 Pedestrian re-identification method and system based on deep learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210599800.4A CN114882441A (en) 2022-05-30 2022-05-30 Pedestrian re-identification method and system based on deep learning

Publications (1)

Publication Number Publication Date
CN114882441A true CN114882441A (en) 2022-08-09

Family

ID=82679255

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210599800.4A Pending CN114882441A (en) 2022-05-30 2022-05-30 Pedestrian re-identification method and system based on deep learning

Country Status (1)

Country Link
CN (1) CN114882441A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117152849A (en) * 2023-10-09 2023-12-01 江苏比特达信息技术有限公司 Novel method for identifying identities of underground weak characteristic personnel of coal mine

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117152849A (en) * 2023-10-09 2023-12-01 江苏比特达信息技术有限公司 Novel method for identifying identities of underground weak characteristic personnel of coal mine

Similar Documents

Publication Publication Date Title
CN107563372B (en) License plate positioning method based on deep learning SSD frame
CN109558823B (en) Vehicle identification method and system for searching images by images
CN108268867B (en) License plate positioning method and device
Quiros et al. A kNN-based approach for the machine vision of character recognition of license plate numbers
CN109034086B (en) Vehicle weight identification method, device and system
CN110781980B (en) Training method of target detection model, target detection method and device
CN111191533A (en) Pedestrian re-identification processing method and device, computer equipment and storage medium
CN111401196A (en) Method, computer device and computer readable storage medium for self-adaptive face clustering in limited space
CN113343985B (en) License plate recognition method and device
CN114519877A (en) Face recognition method, face recognition device, computer equipment and storage medium
CN111291695B (en) Training method and recognition method for recognition model of personnel illegal behaviors and computer equipment
CN113065551A (en) Method for performing image segmentation using a deep neural network model
CN113283414A (en) Pedestrian attribute identification method, related equipment and computer readable storage medium
CN111914668A (en) Pedestrian re-identification method, device and system based on image enhancement technology
CN113963026A (en) Target tracking method and system based on non-local feature fusion and online updating
Parvin et al. Vehicle number plate detection and recognition techniques: a review
CN110991374B (en) Fingerprint singular point detection method based on RCNN
CN114882441A (en) Pedestrian re-identification method and system based on deep learning
CN116597343A (en) Expressway weather identification method and device based on ensemble learning algorithm
CN116152824A (en) Invoice information extraction method and system
CN115731282A (en) Underwater fish weight estimation method and system based on deep learning and electronic equipment
CN116152226A (en) Method for detecting defects of image on inner side of commutator based on fusible feature pyramid
CN111882000A (en) Network structure and method applied to small sample fine-grained learning
CN112784494B (en) Training method of false positive recognition model, target recognition method and device
CN113449676B (en) Pedestrian re-identification method based on two-way interaction-based disentanglement learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination