Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Multi-target detection model construction method based on convolutional neural network

A technology of convolutional neural network and detection model, which is applied in the field of multi-target detection model construction based on convolutional neural network, can solve the problems of poor recognition of complex images or small targets, reduce redundant calculations, improve accuracy, The effect of speeding up calculations

Active Publication Date: 2018-08-17
HENAN UNIVERSITY OF TECHNOLOGY +1
View PDF6 Cites 42 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide a multi-target detection model construction method based on convolutional neural network to solve the problem that the existing detection model is less effective when it is applied to identify complex images or small targets

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
  • Multi-target detection model construction method based on convolutional neural network
  • Multi-target detection model construction method based on convolutional neural network
  • Multi-target detection model construction method based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0027] Example 1: A multi-target detection model based on convolutional neural network, see figure 1 , first input the multi-target image to the shared layer containing 5 convolution layers and 2 max pooling layers for feature extraction, and then input the extracted feature map to ADPN to accurately generate multi-target regions in real time, then The generated multi-object regions are input into DALN to infer the class and location of objects in the region. figure 1 The two output layers of ADPN, Conv_class and Conv_bbr, are output by softmax classifier and bounding box regression respectively. The loss function expression is , ; that is, the weighted sum of the loss function of foreground and background classification and the bounding box regression loss function, and the value of the balance parameter α is set to 2; figure 1 DALN uses two softmax classifier outputs, and its function expression is , that is, the weighted sum of the loss functions of location estimatio...

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 multi-target detection model construction method based on a convolutional neural network to solve the technical problem that the existing detection model cannot distinguish multiple targets and is difficult to identify small targets. The method comprises the following steps: 1: constructing a Caffe deep learning framework, wherein the configuration of the detection modelis completed by using a Faster R-CNN algorithm, and a ZF network is introduced for feature extraction; 2, designing an ADPN network to accurately generate a multi-target area in real time; 3, designing and optimizing the loss function of the ADPN; 4, training the ADPN; 5, designing a DALN sub-network for detecting multi-target categories and locations; 6, designing and optimizing the loss functionof the DALN; 7, training the DALN; and 8, jointly training the ADPN and the DALN to obtain a detection model. The method can identify the targets of multiple categories, improve the recognition ability for the small targets, and is fast in calculation speed and high in precision.

Description

technical field [0001] The invention relates to the technical field of target detection, in particular to a method for constructing a multi-target detection model based on a convolutional neural network. Background technique [0002] Object detection is an important topic in the field of computer vision. The main task is to locate objects of interest from images. It is necessary to accurately determine the specific category of each object and give the bounding box of each object. In recent years, object detection has been widely used in intelligent video surveillance, vehicle autonomous driving, robot environment perception, blind image recognition and other fields. However, object detection is a challenging task due to the deformation of objects due to factors such as perspective, occlusion, and pose. The research focus of traditional object detection methods is on feature extraction and feature classification. As a result, researchers have proposed various forms of featu...

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): G06K9/62G06N3/08G06N3/04
CPCG06N3/08G06N3/045G06F18/241
Inventor 张庆辉万晨霞卞山峰
Owner HENAN UNIVERSITY OF TECHNOLOGY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
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