Method for re-identifying pedestrian with occlusion based on instance segmentation and image restoration

Abstract

The invention discloses a method for re-identifying pedestrian with occlusion based on instance segmentation and image restoration, and the method comprises the steps: obtaining a to-be-searched image, and carrying out the instance segmentation and image restoration of the to-be-searched image, and obtaining a restored pedestrian image; performing instance segmentation on the pedestrian image in the pedestrian image library based on a pre-trained instance segmentation model to obtain a segmented pedestrian image; and performing feature extraction on the repaired pedestrian image and the segmented pedestrian image, performing similarity measurement on the extracted features, and performing retrieval to obtain other images of the same pedestrian. According to the method, pedestrian missing parts are detected and repaired, more complete pedestrian information in a larger area is provided, a subsequent network is helped to obtain more discriminative feature representation, and therefore the pedestrian re-recognition effect is improved. The method for re-identifying the pedestrian with occlusion based on instance segmentation and image restorationn can be widely applied to the field of pedestrian re-identification.

Description

technical field [0001] The invention relates to the field of pedestrian re-identification, in particular to a method for re-identifying pedestrians with occlusions based on instance segmentation and image restoration. Background technique [0002] The task of pedestrian re-identification refers to the task of retrieving the target pedestrian appearing in one camera perspective from other camera perspectives in a multi-camera environment. Although there have been many studies on the general person re-identification work, the person re-identification task for the occlusion problem still has great shortcomings. Some existing occluded pedestrian re-identification methods reduce the impact of occluders on feature extraction by detecting the pixel area where the occluder is located. However, there are problems such as various types and shapes of occluders, and it is impossible to learn occluders well. . Contents of the invention [0003] In order to solve the above-mentioned t...

AI Technical Summary

Problems solved by technology

Although there have been many studies on the general person re-identification work, the person re-identification task for the occlusion problem still has great shortcomings.

Some existing occluded pedestrian re-identification methods reduce the impact of occluders on feature extraction by detecting the pixel area where the occluder is located. However, there are problems such as various types and shapes of occluders, and it is impossible to learn occluders well.

Images