Multi-specialized target detection algorithm based on deep learning

A target detection algorithm and deep learning technology, applied in the field of multi-specialized target detection algorithms, can solve problems such as recognition troubles, and achieve the effect of improving performance and enhancing performance

Active Publication Date: 2019-09-10
NEXWISE INTELLIGENCE CHINA LTD
View PDF2 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, it is rather cumbersome for a single model to accurately identify all appearances, since objects ap

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-specialized target detection algorithm based on deep learning
  • Multi-specialized target detection algorithm based on deep learning
  • Multi-specialized target detection algorithm based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0027] 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.

[0028] refer to figure 1 ,like figure 1 As shown, a multi-specialty target detection algorithm based on deep learning, the detection model framework of the target detection algorithm mainly includes a RoI generation module and a multi-specialty channel module; the data set in it can improve the detection accuracy and robustness of the network model ; The network model can still maintain high detection accuracy in the face of complex data.

[0029] The function of the RoI generation module is: the target detection algorithm obtains an augmented RoI set through multi-scale sliding windows and selec...

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-specialized target detection algorithm based on deep learning. A detection model framework of the target detection algorithm mainly comprises a RoI generation module and a multi-specialized channel module, wherein the target detection algorithm obtains an augmented RoI set through a multi-scale sliding window and selective search, and takes over the multi-scale sliding window in an exhaustion mode to generate a dense RoI set; a plurality of special channel networks are designed, each special channel network is responsible for detecting one type of RoI, and eachspecial network is provided with a full connection layer in the same form. According to the method, two strategies are used for improving the performance, the two strategies enhance the performance ofthe detection model in different aspects, and multiple special channel detection related to the shape category is designed; an augmented RoI dataset is further generated using a selective search andprovided to the network for training.

Description

technical field [0001] The invention relates to a target detection algorithm, in particular to a multi-specialty 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. In general, object detection uses unique shape patterns as evidence to find objects of interest in images. Object detection models are trained on these shape patterns that exhibit classes of different objects. However, it is quite cumbersome for a single model to accurately recognize all the appearances, since objects appear quite differently in images according to their basic shape as well as different poses and viewpoints. Therefore, traditional object detection methods often adopt the method of mixing multiple classifiers, and each classifier is only associated with the corresponding shape pattern in order to better capture the shap...

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): G06K9/62G06N3/04
CPCG06N3/045G06F18/214G06F18/2415
Inventor 龙飞胡建国招继恩王国良段绪海
Owner NEXWISE INTELLIGENCE CHINA LTD
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
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap