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

A People Counting Method Based on Convolutional Neural Network

A convolutional neural network and people counting technology, which is applied in neural learning methods, biological neural network models, neural architectures, etc., to achieve the effect of reducing the accuracy of the method and increasing the universality

Active Publication Date: 2020-09-04
YANSHAN UNIV
View PDF12 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a method for counting people based on convolutional neural network, which can effectively improve the problem of the inconsistency of image scales leading to the decline in the accuracy of counting people through the multi-scale feature fusion network. Context to obtain more contextual information to improve the universality of the people counting method in different scenarios, so as to achieve the purpose of accurate people counting

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 People Counting Method Based on Convolutional Neural Network
  • A People Counting Method Based on Convolutional Neural Network
  • A People Counting Method Based on Convolutional Neural Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0088] figure 1 It is a flow chart of a method for counting people based on a convolutional neural network of the present invention. The method includes the following steps:

[0089] Step 1: Process the image sample data to generate the actual crowd density map:

[0090] 1.1) To grayscale the image in the image sample data, set the three-channel values ​​of the color image to R, G and B, the grayscaled image is Gray, and the grayscale calculation formula is as follows:

[0091] Gray=R*0.299+G*0.587+B*0.114 (1)

[0092] And by dividing the data into image blocks to increase the amount of sample data,

[0093] 1.2) Let any pixel on a single image be x, and the pixel where the head is x g , S is the set of pixels around the person's location, using a two-dimensional Gaussian kernel G with a standard deviation of σ σ And the impulse function δ(x-x g ) Perform convolution* operation to obtain the crowd density map D(x) corresponding to the image. The calculation formula is as follows:

[009...

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 method for counting people based on a convolutional neural network, which relates to the field of computer vision. First, the image sample data is processed and the actual crowd density map of the sample image is generated; then the hierarchical context information is obtained by extracting and processing the feature maps in the branch network and sent to the backbone network, where the low-level and high-level feature maps are selectively fused. , establish a hierarchical context and multi-scale feature fusion network; use the processed sample data to train the established network; finally use the trained model to count the number of people on any image. The invention effectively solves the problem that the accuracy rate decreases due to inconsistent image scales in the people counting task, and improves the universality of the method in different scenarios.

Description

Technical field [0001] The invention relates to the field of computer vision, in particular to a method for counting people based on convolutional neural networks. Background technique [0002] People counting, as one of the most basic and difficult tasks in the fields of crowd analysis, scene understanding, and video surveillance, has attracted widespread attention from academia and industry. People counting is to locate the crowd and estimate the number of crowds by obtaining the density map corresponding to a given image. [0003] At present, people counting methods for a single image are mainly divided into three categories: detection-based methods, regression-based methods and density map-based methods. Detection-based methods and regression-based methods are limited to a certain extent due to phenomena such as severe crowd occlusion and multi-scale performance, while ignoring the key spatial information on the image. Therefore, in recent years, most people counting tasks ha...

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): G06K9/00G06K9/62G06N3/04G06N3/06G06N3/08
CPCG06N3/06G06N3/08G06V20/53G06N3/045G06F18/24G06F18/253
Inventor 张世辉李贺桑榆
Owner YANSHAN 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