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

Image target detection method based on convolutional neural network

A convolutional neural network and target detection technology, applied in neural learning methods, biological neural network models, neural architectures, etc., can solve problems such as easy loss, difficult detection, missed detection, etc., achieve a high degree of automation and meet actual needs , the effect of fast detection speed

Pending Publication Date: 2022-05-17
BEIJING UNIV OF TECH
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In the target detection network, the size of the target to be detected is different, and it is relatively easy to detect a larger target, but for a smaller target, it is easy to lose features in the process of network downsampling, and the detection is difficult. Occurs frequently

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
  • Image target detection method based on convolutional neural network
  • Image target detection method based on convolutional neural network
  • Image target detection method based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] In order to better understand the technical solution of the present invention, the implementation of the present invention will be further described in detail below.

[0024] The present invention is implemented under the framework of Pytorch by using Python language programming. First complete the construction of the network and configure the relevant initial parameters; use the trained network model to detect the target on the image.

[0025] Dataset preprocessing. First, use the data labeling tool to mark the image target, and divide all the data into a training set and a verification set in a ratio of 4:1. The training set is used to train the model, and the verification set is used to test the model performance parameters after each training epoch. Using data enhancement technology on the training set, Mosaic data enhancement first randomly reads four pictures from the training set, performs operations such as inversion, scaling, and color gamut changes on the fo...

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 an image target detection method based on a convolutional neural network, and the method comprises the following steps: carrying out the preprocessing of a data set, and employing a data enhancement method; a Mosaic data enhancement method is used for data, and four random pictures are spliced together to improve the data training effect; a basic convolutional neural network is constructed by referring to YOLOv2, each convolution unit comprises a convolution layer, a BatchNormalization layer and a Relu activation layer, no full connection layer exists in the whole network structure, and the convolution layers are completely used; a residual network structure is used in the convolutional neural network, so that the depth of the network can be deepened, and a better feature is provided for model learning; multi-scale feature map fusion is used to improve the detection effect of a network model on images of different scales, and the robustness of a target detection model is improved; data annotation is made after data enhancement, then clustering analysis is performed on bounding boxes in a training set by using a clustering method, and a proper prior box is selected according to a clustering analysis result, so that model learning is facilitated.

Description

technical field [0001] The invention belongs to the fields of artificial intelligence, deep learning and computer vision, and is an image target detection method based on a convolutional neural network. It draws on some experiences in the field of target detection and builds a convolutional neural network to provide an image target detection method based on a convolutional neural network. Methods in the field of image object detection. Background technique [0002] Image object detection is an important application of artificial intelligence in the field of computer vision. Now the deep learning model has replaced the traditional machine vision method and has become the mainstream solution in the field of target detection. Generally speaking, target detection usually focuses on a specific object target, and requires both the category information and location information of this target. We need to separate the target of interest from the background and determine the descript...

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 Applications(China)
IPC IPC(8): G06V10/44G06V10/764G06V10/82G06V10/80G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/048G06N3/045G06F18/241G06F18/253
Inventor 李永王学舟
Owner BEIJING UNIV OF 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