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

Crowd counting method and system based on multi-scale interactive network

A crowd counting, multi-scale technology, applied in neural learning methods, biological neural network models, computing, etc.

Pending Publication Date: 2021-05-07
SHANDONG NORMAL UNIV
View PDF8 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The inventor found that although great progress has been made so far, the scale change of the crowd in an image is continuous, and only discrete-scale crowd features can be extracted through convolution kernels of different sizes
Therefore, how to extract more effective information from images with varying scales is still a challenging problem.

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
  • Crowd counting method and system based on multi-scale interactive network
  • Crowd counting method and system based on multi-scale interactive network
  • Crowd counting method and system based on multi-scale interactive network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] The purpose of this embodiment is to provide a crowd counting method based on multi-scale interaction.

[0041] A crowd counting method based on multi-scale interaction, including:

[0042] Obtain the image of the crowd to be counted;

[0043] Inputting the image of the crowd to be counted into the pre-trained multi-scale interactive network, and outputting the crowd density map; adding and summing the pixel values ​​of the density map to obtain the final crowd counting result;

[0044] Wherein, the multi-scale interactive network includes an encoder, an aggregation interaction module and a decoder, the encoder includes a multi-layer feature extraction module, and the multi-layer feature extraction module extracts multi-layer features through VGG-16, and the aggregation interaction The modules effectively fuse features of adjacent scales by learning from each other, and the decoder includes a self-interaction module that obtains rich scale information from the extracte...

Embodiment 2

[0066] The purpose of this embodiment is to provide a crowd counting system based on multi-scale interaction.

[0067] A crowd counting system based on multi-scale interaction, including:

[0068] an image acquisition unit configured to acquire images of people to be counted;

[0069] The counting unit is configured to input the image of the crowd to be counted into the pre-trained multi-scale interactive network, and output the crowd density map; the pixel values ​​of the density map are accumulated and summed to obtain the final crowd counting result;

[0070] Wherein, the multi-scale interactive network includes an encoder, an aggregation interaction module and a decoder, the encoder includes a multi-layer feature extraction module, and the multi-layer feature extraction module extracts multi-layer features through VGG-16, and the aggregation interaction The modules effectively fuse features of adjacent scales by learning from each other, and the decoder includes a self-in...

Embodiment 3

[0072] The purpose of this embodiment is to provide an electronic device.

[0073] An electronic device, including a memory, a processor, and a computer program stored on the memory, when the processor executes the program, the multi-scale interactive crowd counting method is implemented, including:

[0074] Obtain the image of the crowd to be counted;

[0075] Inputting the image of the crowd to be counted into the pre-trained multi-scale interactive network, and outputting the crowd density map; adding and summing the pixel values ​​of the density map to obtain the final crowd counting result;

[0076] Wherein, the multi-scale interactive network includes an encoder, an aggregation interaction module and a decoder, the encoder includes a multi-layer feature extraction module, and the multi-layer feature extraction module extracts multi-layer features through VGG-16, and the aggregation interaction The modules effectively fuse features of adjacent scales by learning from eac...

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 provides a crowd counting method and system based on multi-scale interaction, and the method and system effectively meet the multi-scale requirements of a crowd in a timely task through a multi-scale interactive network. In the network, the aggregation interaction module can effectively utilize context features of adjacent layers in a mutual learning mode, and the self-interaction module enables the network to adaptively extract multi-scale information from images and better cope with continuous changes of crowd scales in the images.

Description

technical field [0001] The present disclosure relates to the technical field of computer vision image processing, in particular to a multi-scale interactive crowd counting method and system. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] Crowd counting occupies a very important and indispensable position in the field of social security protection. Since mass gatherings are ubiquitous and necessarily frequent at sporting, musical, political and other social events. Automated crowd counting plays an important role in the processing and analysis of such events. Crowd counting is an active research area in computer vision because many key issues have not been reasonably addressed, such as severe occlusions, huge crowd diversity within and across different regions in an image, and large viewing angle variations, especially at scale Continu...

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): G06T7/00G06N3/04G06N3/08
CPCG06N3/04G06N3/08G06T7/00G06T2207/20081G06T2207/20084G06T2207/30196G06T2207/30242
Inventor 李皓
Owner SHANDONG NORMAL UNIV
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