Pedestrian Re-Identification Method Based on Visual Saliency and Multiple Features
A pedestrian re-identification, multi-feature technology, applied in the field of image processing, can solve the problems of low pedestrian re-recognition accuracy, pedestrian targets are easily affected by illumination, perspective, and posture changes, and achieve the effect of improving accuracy.
Active Publication Date: 2018-12-18
江苏睿世力科技有限公司
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0007] The technical problem to be solved by the present invention is: pedestrian targets in the monitoring network are easily affected by illumination, viewing angle, and posture changes, resulting in low pedestrian re-identification accuracy
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
[0019] Relevant detailed content and technology of the present invention, hereby cooperate drawing description as follows:
[0020] The implementation schematic diagram of the present invention is as figure 1 As shown, the specific steps are as follows:
[0021] (1) Detect pedestrians in the video, use the visual saliency model to preprocess the detected pedestrian result frame, and segment the complete pedestrian target;
[0022] In step (1), the saliency map M of the pedestrian area is obtained and the target segmentation is performed, and the specific process includes:
[0023] Calculate the original grayscale image M of the region of interest 0 and binary map M 1 , combining the two to generate a continuous and smooth weakly saliency map M w ,Calculated as follows:
[0024] (Formula 1)
[0025] According to the weak saliency map M w Select a training set containing positive and negative samples, and use the Multiple Kernel Boosting method to train the samples to ...
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 pedestrian re-identification method based on visual salience and multiple features. The method includes: fast detection of pedestrian targets in the video; segmenting the pedestrian area using a saliency detection algorithm, and extracting the spatial histogram of the target area; using the SCA method to segment the human body into three parts: head, torso, and legs, and extracting the torso and legs respectively The color, position, shape and texture features of the object; the improved JSD similarity measurement criterion is used to calculate the similarity of the spatial histogram for initial target recognition; the Gaussian function is used to calculate the similarity of the regional feature set to obtain the final recognition result. The method provided by the invention can realize long-term tracking and monitoring of specific pedestrians in a multi-camera network under different background environments and camera settings. Realize the intelligent processing of surveillance video, respond to abnormal conditions in the video in a timely manner, and save a lot of manpower and material resources.
Description
technical field [0001] The invention belongs to the technical field of image processing, and relates to computer vision and pattern recognition methods, in particular to a method for realizing pedestrian re-identification. Background technique [0002] With the advent of the era of the Internet of Things, the country vigorously advocates the construction of smart cities, and topics based on video analysis such as intelligent security and intelligent transportation are continuously proposed, and more and more attention is paid to the intelligent analysis of acquired surveillance videos. For most monitoring systems, the people appearing in the monitoring video are the important objects of attention, so the intelligent monitoring system needs to have the ability to detect, identify and track them, so as to further analyze the behavior of pedestrian targets, pedestrians Re-identification technology is one of its core steps. The so-called pedestrian re-identification is to re-id...
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): G06K9/00
CPCG06V20/53
Inventor 仇春春李庆武周亮基程海粟
Owner 江苏睿世力科技有限公司