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

Target detection method based on lightweight backbone network

A backbone network and target detection technology, applied in the field of target detection, can solve problems such as low detection efficiency and inaccurate positioning of bounding boxes, and achieve the effects of improving accuracy, improving detection performance, and ensuring detection accuracy.

Pending Publication Date: 2022-08-05
XIAN UNIV OF TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide a target detection method based on a lightweight backbone network, which solves the problems of low detection efficiency and inaccurate bounding box positioning in existing detection technologies

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
  • Target detection method based on lightweight backbone network
  • Target detection method based on lightweight backbone network
  • Target detection method based on lightweight backbone network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0064] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0065] The present invention provides a target detection method based on a lightweight backbone network, such as Figure 1-3 shown, follow the steps below:

[0066] Step 1. Prepare the pre-training dataset Tiny-Imagenet and the training dataset in VOC format;

[0067] Step 1 is implemented according to the following steps:

[0068] Step 1.1, prepare the pre-training dataset Tiny-Imagenet, and generate the label files required for pre-training;

[0069] Step 1.1 is implemented according to the following steps:

[0070] Step 1.1.1. Extract 3.1G from the Imagenet data set to make the required pre-training data set Tiny-Imagenet (TI); since Imagenet is a very large open source data set (about 100G or so), perform it directly on Imagenet Training can take a long time and is very hardware-intensive, so we sampled a small subset of Imagenet to mak...

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 target detection method based on a lightweight backbone network. The target detection method is specifically implemented according to the following steps: step 1, preparing a pre-training data set Tiny-Image and a training data set in a VOC format; step 2, establishing a feature extraction network LBNet; step 3, constructing a feature pyramid structure formed by multilayer output features of the LBNet, and performing pre-training in the Tiny-Image data set obtained in the step 1 to obtain a pre-training weight; and step 4, designing an anchor generation container algorithm, and generating an anchor size required by the region recommendation network. And 5, completing the training and prediction of the model. According to the method, the problems of low detection efficiency and inaccurate bounding box positioning in the existing detection technology are solved.

Description

technical field [0001] The invention belongs to the technical field of target detection, and relates to a target detection method based on a lightweight backbone network. Background technique [0002] The task of object detection is mainly to identify all object categories in a given input image and determine their bounding box locations. With the vigorous development of deep learning, object detection technology has been widely used in autonomous driving, face recognition and other fields. However, most feature extraction networks currently used for target detection models have complex structures and large number of parameters, which leads to the problem of low detection efficiency; secondly, most of the existing detection models are based on the work experience of researchers for regional recommendation networks. Manually preset rectangular bounding boxes (anchors) are provided, and these anchors have a low degree of fitting with the experimental dataset, which greatly af...

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): G06V10/774G06V10/40G06V10/77G06V10/762G06V10/82G06N3/04G06N3/08
CPCG06V10/774G06V10/40G06V10/7715G06V10/762G06V10/82G06N3/08G06V2201/07G06N3/048G06N3/045
Inventor 刘晶马小龙梁庆国
Owner XIAN UNIV OF TECH