Video re-identification training method, re-identification method and storage device
A training method and re-identification technology, applied in the field of computer vision, can solve the problem of insufficient pedestrian re-identification accuracy, and achieve the effect of improving accuracy and robustness
Active Publication Date: 2020-11-10
ZHEJIANG DAHUA TECH CO LTD
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
These methods usually only focus on the global features of the pedestrian sequence, and do not focus on the salient features of the pedestrian's face, torso and other key parts, resulting in insufficient accuracy of video-based pedestrian re-identification
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.
[0025] The terms "first", "second", and "third" in this application are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of indicated technical features. Thus, features defined as "first", "second", and "third" may explicitly or implicitly include at least one of these features. In the description of the present application, "plurality" means at least t...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a video re-identification training method, a re-identification method and a storage device. The re-identification training method includes: using animal detection and animal tracking methods to detect animal picture sequences in video; extracting the time domain of animal picture sequences Features and spatial features, integrate temporal features and spatial features to obtain feature maps of animal picture sequences; perform block processing on feature maps of different sizes in the horizontal dimension, and calculate local block feature maps and global feature maps respectively The loss between real animals; optimize this loss for training until the training converges to obtain the optimal animal re-identification result. Through the above method, the present invention can not only fuse the time-domain features and spatial features of animal picture sequences, but also perform fine-grained learning on different parts of animals while taking into account the learning of global features, improving the accuracy and robustness of animal re-identification sex.
Description
technical field [0001] The present application relates to the technical field of computer vision, in particular to a video re-identification training method, a re-identification method and a storage device. Background technique [0002] Surveillance video is often used in subways, airports, traffic roads and other public places to maintain public safety, detect pedestrians through video and use pedestrian re-identification technology to find suspects or lost children. Traditional pedestrian re-identification methods mostly use a single pedestrian image for retrieval, but in surveillance videos, the posture, occlusion, and environmental background of pedestrians may vary with time. The robustness of single image retrieval results Not strong. However, the video-based pedestrian re-identification method uses multiple pictures in the video sequence for recognition, and the recognition effect is better. [0003] The current method to solve the problem of video pedestrian re-ide...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06N3/04
CPCG06V40/25G06V40/10G06V20/40G06V20/52G06N3/045
Inventor 张迪潘华东罗时现
Owner ZHEJIANG DAHUA TECH CO LTD