A real-time online multi-target tracking method using temporal dynamic appearance model
A multi-target tracking and dynamic table technology, applied in character and pattern recognition, instrumentation, computing, etc., can solve the problems of spatial distribution confusion, limiting the role of apparent models, etc., and achieve the effect of high computing efficiency
Inactive Publication Date: 2019-04-05
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the task of multi-target tracking requires the appearance model to be able to distinguish multiple targets belonging to the same category. It is easy to cause confusion only by relying on the spatial distribution of appearance features, which limits the role of the appearance model in practice.
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
[0021] The invention is a real-time online multi-target tracking method of a temporal dynamic appearance model. By considering the change rule of the target's appearance information over time during the tracking process, an appearance model with strong discriminative power is established to provide accurate appearance matching information for the data association between the target and the detection results, and the detection results are frame by frame The task of online multi-target tracking is completed by associating with the target accurately.
[0022] The following describes the implementation of the present invention in detail through specific examples. A flow chart of a real-time online multi-target tracking method using a temporal dynamic appearance model is as follows figure 1 As shown, the implementation process is as follows:
[0023] According to step 1: use the target detector trained offline for each frame of video image to detect the target, and obtain the dete...
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 relates to a real-time online multi-target tracking method using a temporal dynamic appearance model. The present invention regards the appearance information of each target in the process of multi-target tracking as a time-ordered appearance sequence, and establishes a table with strong discriminative power by considering the rule of time-varying target appearance information in the tracking process. At the same time, multiple hidden state observation models are used to model the structural diversity of the target appearance in the feature space, providing an accurate representation of the data association between the target and the detection results. View matching information; by associating the detection results with the target frame by frame to complete the task of real-time online multi-target tracking, it can accurately track multiple targets of interest in the video in real time. It has high computing efficiency while ensuring tracking performance, and can be practically applied to various real-time systems.
Description
technical field [0001] The invention relates to a real-time online multi-target tracking method using a temporal dynamic appearance model, which is used for real-time online positioning of a target of interest in a video and real-time recovery of its motion track, and belongs to the technical fields of computer vision and video monitoring. Background technique [0002] Visual multi-object tracking aims to localize, recognize multiple objects in a scene from video, and estimate the trajectory of each object. Multiple object tracking is the basis of many high-level tasks in computer vision, such as scene understanding, event detection, action recognition, etc. Multi-target tracking is also widely used in systems such as video surveillance, human-computer interaction, and visual navigation. [0003] The detection-based multi-target tracking method is one of the mainstream methods of visual multi-target tracking. Its main idea is to train detectors for specific categories of ta...
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): G06K9/62
CPCG06F18/24
Inventor 贾云得杨敏裴明涛
Owner BEIJING INSTITUTE OF TECHNOLOGYGY