Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A Fast Multi-Scale Estimation Object Tracking Method on Re-Detection

A re-detection and target tracking technology, which is applied in the fields of image processing and computer vision, can solve problems such as error and slow tracking speed, and achieve the effect of solving fast motion, improving computing speed, and improving feature expression ability

Active Publication Date: 2022-06-07
NANJING UNIV OF INFORMATION SCI & TECH
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although the LCT algorithm improves the robustness in the case of target occlusion to a certain extent, there is still a problem. It only judges whether re-detection is required by whether the maximum response value is lower than the threshold value. This method of discrimination will have a large error.
In addition, because it is necessary to build a pyramid model on the target to estimate the optimal scale of the target, this will make the overall tracking rate slower

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Fast Multi-Scale Estimation Object Tracking Method on Re-Detection
  • A Fast Multi-Scale Estimation Object Tracking Method on Re-Detection
  • A Fast Multi-Scale Estimation Object Tracking Method on Re-Detection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0069] The system flow frame diagram of re-detected target tracking is as follows figure 1 As shown, it includes the following parts: training filter template, localization, detection, re-detection, and model update.

[0070] (1) Step 1: Train the filter template. First, initialize the target tracker, mark the initial area of ​​the target, use the VGG-19 network to extract the depth features of the target, and establish the initial target template and scale template for the calculation of the target response value in the second frame.

[0071] The establishment of the target template is mainly divided into the following parts:

[0072] First, the classifier performs cyclic shift sampling on an image block of size M×N with the target position as the center, and the resulting sample set is denoted as x i , where i∈{0,…M-1}×{0,…N-1}. each sample x i has a corresponding regression label y i , y i obtained from the Gaussian function. The purpose of the classifier f(x) is to ...

example

[0119] The invention measures the performance of the tracking algorithm through the OPE (one pass evaluation) evaluation standard, and selects 60 challenging video sequences from the OTB100 data set for analysis, and compares it with other trackers (DeepKCF, SAMF, KCF, CSK, DFT) , CT, CACF and other 7 trackers) are compared under different challenge factors (illumination change, target deformation, motion blur, fast motion, in-plane rotation, out-of-plane rotation, target out of field of view, background clutter, low resolution, etc.) .

[0120] Figure 4 are the sampling frames of the tracking results of the tracking method of the present invention (DRKCF) ​​and other seven trackers, from Figure 4 It can be seen that the tracker proposed by the present invention can track the target better than other trackers, and even if the target is lost, the target can be retrieved to continue tracking.

[0121] Figure 5 is the comparison of the tracking method of the present inventi...

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

No PUM Login to View More

Abstract

The present invention proposes a fast multi-scale estimation target tracking algorithm on deep features and re-detection. The characteristics of the target are represented by the method of deep learning, which improves the feature expression ability of the target. In the tracking stage, when extracting features of image blocks of different scales, PCA dimensionality reduction can reduce the amount of calculation and improve the overall calculation speed. Based on two discriminant indicators, peak side lobe ratio (PSR) and confidence smoothing constraint (SCCM), a new detection index is proposed, so that the tracking reliability of the current frame can be measured more accurately. If the reliability of the current frame is low, a series of target candidate boxes are generated by the method of Edgeboxes for re-detection.

Description

technical field [0001] The invention belongs to the fields of image processing and computer vision, and uses a deep learning method to learn target features, and realizes accurate tracking of the target by re-detecting when the target drifts. It can be used in areas such as unmanned driving and video surveillance. Background technique [0002] Object tracking is a key problem in computer vision, and has a wide range of applications in various fields such as video surveillance, behavior recognition, unmanned driving, and medical images. The purpose of target tracking is to estimate the target position for each subsequent frame given the initial position of the target in the first frame. At present, the main computer vision tracking methods mainly include the tracking method based on correlation filtering and the tracking method based on deep learning. [0003] Target tracking algorithms based on correlation filtering have developed rapidly since 2010, among which Henriques ...

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
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06V10/77G06K9/62G06V10/764G06V10/774
CPCG06F18/2135G06F18/2411G06F18/214
Inventor 胡昭华黄嘉净
Owner NANJING UNIV OF INFORMATION SCI & TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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