Pedestrian re-identification method based on spatial reverse attention network
A pedestrian re-identification, space technology, applied in neural learning methods, character and pattern recognition, biological neural network models, etc., to achieve the effect of improving effectiveness and reliability
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0039] The task of person re-identification aims to find the same pedestrian under different cameras. Although the development of deep learning has brought great improvements to person re-identification, it is still a challenging task. In recent years, attention mechanisms have been widely verified to have excellent effects on person re-identification tasks, but the combined effects of different types of attention mechanisms (such as spatial attention, self-attention, etc.) still remain to be discovered.
[0040] refer to Figure 1~3 , as an embodiment of the present invention, provides a method for pedestrian re-identification based on spatial reverse attention network, including:
[0041] S1: Collect the captured pictures and divide them into training set and test set;
[0042]S2: Construct a spatial reverse attention network model based on Resnet-50, train the convolutional neural network according to the training set, and add CBAM-Pro;
[0043] It should be noted that th...
Embodiment 2
[0083] In order to verify and explain the technical effect adopted in this method, this embodiment adopts the traditional technical scheme and the method of the present invention to conduct a comparative test, and compares the test results by means of scientific demonstration to verify the real effect of this method.
[0084] In this embodiment, experiments are carried out on the three most commonly used data sets for pedestrian re-identification tasks: Marker-1501, DukeMTMC-reID, and CUHK03, using the first successful matching probability (rank-1) and average precision (mean average precision, mAP ) to evaluate the experimental results.
[0085] Among them, Marker-1501 includes 1501 pedestrians with different identities captured by 6 cameras. The data set generates 32668 pictures containing individual pedestrians through the DPM detector. They are divided into non-overlapping training / testing sets. The training set Contains 12,936 images of 751 pedestrians with different iden...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com