A Robust Convolution Kernel Number Adaptation Method Based on Corner Radiation Domain

A convolution kernel and radiation area technology, applied in the field of convolution kernel quantity adaptation, can solve the problems of convolution kernel size limitation, poor interpretability, high computing overhead, etc., to reduce computing cost, high interpretability, and improve accuracy rate effect

Active Publication Date: 2021-06-18
SHANDONG UNIV OF SCI & TECH
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, existing methods for adapting the number of convolution kernels are limited by the size of the convolution kernels, resulting in problems such as poor interpretability and high computational overhead.

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 Robust Convolution Kernel Number Adaptation Method Based on Corner Radiation Domain
  • A Robust Convolution Kernel Number Adaptation Method Based on Corner Radiation Domain
  • A Robust Convolution Kernel Number Adaptation Method Based on Corner Radiation Domain

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0050] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0051] In convolutional neural network (CNN), the convolutionary core mainly performs convolution, and the convolutionary core is typically initialized in the form of a random matrix, and it is learned by the error reverse propagation (BP) during the network training. After BP operation, convolution kernels continue to learn valuable features to increase their weight. For ease of classification, CNN will bring superimposed weight into the score function. The larger the activation value obtained by the score function, the more requiring the volume of the volume, and the easier the target image is separated from the other image.

[0052] The convolver is like a filter for extracting local features of the image. This model can only be extracted when the convolution layer contains only one convolution. Obviously, the eigenvalues ​​extracted by a single convolutionary core cannot be classified. Therefore, more features can be l...

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 present invention provides a method for adapting the number of robust convolution kernels based on the corner radiation domain, which can adaptively determine the number of convolution kernels, by sequentially performing denoising operations, corner detection operations and corner radiation domain operations , the denoising operation and the corner detection operation are used to extract points with rich features in the image and facilitate the extraction in the corner radiation domain, and the corner radiation domain operation is used to find the corner extension area of ​​the auxiliary function. The method of the present invention is not limited by the size of the convolution kernel, and the output results are consistent and robust for different sizes. The convolution kernel adapted by this method matches the specific data set well, and has good performance on each data set. Compared with the adaptation method with comparable calculation cost, this method can improve the accuracy rate by 3%. More importantly, Compared with the adaptation method with similar accuracy, this method can reduce the calculation cost by 15%.

Description

Technical field [0001] The present invention relates to the field of conjunction, and more particularly to a rack of robs based on a robbier volume nucleus. Background technique [0002] In recent years, deep learning has attracted extensive attention from all walks of life, promoting rapid development of a series of applications in artificial intelligence. As an important branch of deep learning, convolutional neural network (CNN) is closely related to computer vision research. As the CNN model is invested in the computer visual field, computer visual has made excellent progress. The characteristic of convolutional neural network has been widely successful in a series of fields such as computer vision, and gradually replaces traditional artificial design characteristics, becoming new research hotspots. [0003] As one of the important hyperfeit parameters in the CNN, the number of convolutionary cores determines the accuracy of the runtime, storage cost, and training model of th...

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 Patents(China)
IPC IPC(8): G06K9/62G06K9/40G06K9/46
CPCG06V10/30G06V10/44G06F18/214
Inventor 杜玉越梁煜王路张福新亓亮刘伟
Owner SHANDONG UNIV OF SCI & TECH
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