Unlock instant, AI-driven research and patent intelligence for your innovation.

Vehicle re-identification method based on multi-angle deep convolutional neural network

A deep convolution and neural network technology, applied in the field of vehicle re-identification of deep convolutional neural network, can solve problems such as high complexity, improve accuracy, correct redundant and misleading problems, and eliminate redundancy and the effect of misleading questions

Active Publication Date: 2019-11-01
JILIN UNIV
View PDF7 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the complexity of training is large, and if the number of samples is small, there will be over-fitting problems

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
  • Vehicle re-identification method based on multi-angle deep convolutional neural network
  • Vehicle re-identification method based on multi-angle deep convolutional neural network
  • Vehicle re-identification method based on multi-angle deep convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0026] as attached figure 1 The figure shown intuitively shows the meaning of the vehicle re-identification task. The essence of vehicle re-identification is the re-identification of the target, which is to find image samples belonging to the same target as the specified target image from a data set. That is, vehicle re-identification is the process of matching and finding out the same vehicle as the specified vehicle image in multiple surveillance video frames. It can distinguish very well that in the eyes of human vision, it may not feel the same car due to lighting, angle and other issues, but it is actually the same car.

[0027] The present invention performs vehicle re-identification under the VeRi data set, and adjusts the sample size under this example to a size of 128×128 pixels.

[0028]Dataset: As shown in Figure 2, training data and test data are required for vehicle re-identification. Figure 2(a) shows some examples in the training set, and Figure 2(b) shows some...

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 invention discloses a vehicle re-identification method based on a multi-angle deep convolutional neural network. The method comprises the following steps: taking a vehicle image data set as input,and outputting a processed feature image set through a sharing layer, wherein the feature image set is classified from three perspectives of measurement, vision and attributes through three differentstreams; and performing joint learning on the three classification results to realize vehicle re-identification, wherein the three different flows are a cluster-based triad flow, a complementary learning-based appearance flow and a vehicle attribute-based attribute flow. According to the method, the problems of redundancy and misleading caused by random triple sampling are solved. The problem that similar images cannot be well distinguished or the network is not convergent in an existing method is effectively solved. The training time cost is reduced, and a more accurate vehicle re-identification result can be obtained.

Description

technical field [0001] The present invention relates to a search in a database for images containing the same vehicle captured by multiple cameras, and in particular provides a vehicle re-identification method based on a multi-angle deep convolutional neural network; it belongs to public security, large-scale monitoring The technical field of vehicle search and re-identification in image and video databases. Background technique [0002] The use of surveillance cameras in the field of public safety has exploded, and vehicles, as important objects in urban surveillance, have attracted extensive attention in a large number of vehicle-related tasks such as detection, tracking, classification, and verification. Vehicle re-identification is to find out images captured by other cameras that contain the same vehicle as the query image or vehicles under different lighting and viewing angles under the same camera. Through vehicle re-identification, target vehicles can be automatical...

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
IPC IPC(8): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06V20/584G06V2201/08G06N3/045G06F18/22
Inventor 梁艳花秦贵和邹密晏婕赵睿许骞艺张钟翰成一铭
Owner JILIN UNIV
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com