Low-resolution pedestrian re-identification system and method based on super-resolution image generation
A pedestrian re-identification and super-resolution image technology, which is applied in the field of low-resolution pedestrian re-identification system, can solve the problems of poor pedestrian re-identification effect and low-resolution loss, so as to improve the recognition rate of low-resolution pedestrian re-identification and expand the capacity Effect
Active Publication Date: 2021-05-25
SUN YAT SEN UNIV
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0009] Aiming at the defects and deficiencies of the existing technology, and aiming at the problem that pedestrian re-identification is poor due to the loss of discriminative information due to the low resolution of pedestrian images, the present invention provides a low-resolution pedestrian re-identification based on super-resolution images. Recognition system and method to improve the effect of low-resolution pedestrian re-identification by jointly training pedestrian attribute-guided discrimination pedestrian image generation network model and pedestrian re-identification network model
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
[0097] Aiming at the low-resolution pedestrian re-identification problem, this embodiment proposes a low-resolution pedestrian re-identification system based on super-resolution image generation, which includes a pedestrian attribute-guided super-resolution image generation network model and a pedestrian re-identification network model ,
[0098] In this embodiment, the super-resolution image generation network model guided by pedestrian attributes includes a generator and a discriminator, the generator uses a conditional loop against the generator in the generation network, and the discriminator uses a conditional loop against the discriminator in the generation network; The generator includes the first generator G h→l and the second generator G l→h , the discriminator includes a first discriminator D h→l , the second discriminator D l→h ; the first generator G h→l Used to generate low-resolution images from high-resolution images, the second generator G l→h Used to gene...
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 low-resolution pedestrian re-identification system and method based on super-resolution image generation. The system includes a pedestrian attribute-guided super-resolution image generation network model and a pedestrian re-identification network model. Low-resolution image samples and pedestrian attribute vectors; training super-resolution image generation network model; training pedestrian re-identification network model; joint training super-resolution image generation network model and pedestrian re-identification network model; low-resolution pedestrian image test set and The corresponding pedestrian attribute vector is input to the super-resolution image generation network model and the pedestrian re-identification network model after joint training to extract the pedestrian image features; the cosine similarity of the pedestrian image features is calculated, and the matching of pedestrian images with different resolutions is obtained according to the cosine similarity. As a result, the present invention realizes the restoration of low-resolution image details, expands the capacity of the network at the same time, and improves the effect of low-resolution pedestrian re-identification.
Description
technical field [0001] The invention relates to the fields of computer vision and intelligent monitoring, in particular to a low-resolution pedestrian re-identification system and method based on super-resolution image generation. Background technique [0002] The task of pedestrian re-identification is to identify pedestrian identity information across cameras. With the rapid development of video surveillance, pedestrian re-identification plays an increasingly important role in intelligent surveillance. For the problem of pedestrian re-identification, the usual steps are to extract the features of pedestrians first, measure the distance of the features of pedestrians to obtain similarity scores, and then sort according to the similarity scores, and finally obtain the identity of pedestrians. In real life, due to the different positions of the camera and the change of the distance between the pedestrian and the camera, the resolution of the pedestrian images acquired in thi...
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): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V40/10G06N3/045G06F18/24
Inventor 郑伟诗叫洁宁
Owner SUN YAT SEN UNIV