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

A multi-scale conversion target detection algorithm based on deep learning

A target detection algorithm and deep learning technology, applied in the field of multi-scale conversion target detection algorithm, to achieve easy training, improve the effect of information flow and gradient

Inactive Publication Date: 2019-06-25
NEXWISE INTELLIGENCE CHINA LTD +1
View PDF3 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in order to improve the detection performance, the architecture of the above methods must be carefully constructed by adding many convolutional and pooling layers. the cost of

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 multi-scale conversion target detection algorithm based on deep learning
  • A multi-scale conversion target detection algorithm based on deep learning
  • A multi-scale conversion target detection algorithm based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035] In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

[0036] refer to figure 1 ,Such as figure 1 As shown, a multi-scale conversion target detection algorithm based on deep learning includes the following steps:

[0037] S1, using the base network for feature extraction; the base network as a network architecture directly connects all layers with matching map feature sizes, each layer takes additional input from all previous layers, and maps its own Features are passed to all subsequent layers; instead of combining features by summing them before being passed to the next network layer, they are combined by concatenating them;

[0038] S2, using a m...

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-scale conversion target detection algorithm based on deep learning, and the algorithm is characterized in that the algorithm comprises the following steps: S1, carryingout the feature extraction through employing a basic network; S2, adopting a multi-scale conversion module to process the features generated by the convolutional network; And S3, carrying out accurate positioning on the target and classifying the output target by adopting a target positioning and classifying module. According to the algorithm, the information flow and the gradient of the whole network are improved, so that the training is easier. Each layer can directly access a loss function and a gradient of an original input signal, thereby realizing implicit deep monitoring, and helping to train a network architecture more deeply.

Description

technical field [0001] The invention relates to a multi-scale conversion target detection algorithm, in particular to a multi-scale conversion target detection algorithm based on deep learning. Background technique [0002] At present, object detection, as one of the classic research contents in computer vision, has received more and more attention in the research field. Target detection is to identify the category of the target from the background information by analyzing the scene video image frame obtained by the imaging sensor, and give the position information of the target in the image, so as to provide reliable information for subsequent target tracking, scene recognition and other tasks. Data Sources. Therefore, target detection is also widely used in various fields, such as transportation, intelligent security, military fields and so on. In recent years, convolutional neural networks have achieved great success in computer vision tasks such as object detection. T...

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): G06K9/46G06K9/62
Inventor 招继恩龙飞胡建国杨焕朱勇杰王国良
Owner NEXWISE INTELLIGENCE CHINA LTD
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