Ice sublayer structure extraction method based on multi-scale attention mechanism

A multi-scale, attention-based technology, applied in neural learning methods, neural architectures, computer components, etc., to achieve the effect of improving accuracy

Pending Publication Date: 2021-07-09
BEIJING UNIV OF TECH
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, the task of analyzing topological slices of ice sheet radar is more challenging and meaningful

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
  • Ice sublayer structure extraction method based on multi-scale attention mechanism
  • Ice sublayer structure extraction method based on multi-scale attention mechanism
  • Ice sublayer structure extraction method based on multi-scale attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0031] The specific implementation method of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0032] 1. Input data processing

[0033] Arrange the radar topology sequence of consecutive T frames in order, where T is 5, and the data in the shape of 1×5×64×64 (the number of channels×the number of slices×the height of the radar slice map×the width of the radar slice map) is obtained Ready to enter the network.

[0034] 2. Build the MsANet network

[0035] like figure 1 shown. The specific parameters of each layer of the constructed MsANet network of the present invention are as follows:

[0036]① Block 1: 3D convolution unit, 3D batch normalization layer, Relu activation function and hybrid pooling layer are arranged in order. 3D convolution unit: the input size is 5×64×64, the number of input channels is 1, the convolution kernel is 3×5×3, the step size is 1, the edge padding method is “zero padding”, and the output si...

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 ice sublayer structure extraction method based on a multi-scale attention mechanism, and relates to the field of computer vision. The invention provides a novel MsANet network, the network takes a double-branch convolutional network as a backbone network, not only fuses multi-scale features, but also uses a 3D attention mechanism to further carry out additional feature modeling on unique features of different ice layers in a 3D radar topological sequence, so that refining processing of spatial relationships of different ice layers is realized. The attention multi-scale module formed by the 3D attention mechanism and the multi-scale module enables detected important ice layer features to obtain richer scale features by using the multi-scale module, and further strengthens the modeling capability of the key ice layer features. According to the method, the positions of a plurality of ice layers are detected at the same time as different tasks, the unique features of the ice layers at different positions are learned through the two branch structures respectively, and finally a rapid and high-precision ice layer structure extraction algorithm based on the MsANet network is achieved.

Description

technical field [0001] The invention belongs to the fields of computer vision, pattern recognition and polar glaciology, and designs a subglacial structure extraction method based on MsANet network. Background technique [0002] With the improvement of people's living standards and the development of technology, the expectation and demand for intelligent analysis and manufacturing are becoming stronger. Among them, big data analysis and artificial intelligence, which are mainly involved, have received extensive attention. Now, the combination of artificial intelligence and different fields is a direction that is constantly being explored and is of great significance. Polar glaciology, as a complex subject of geographic observation and modeling, is of great significance to human production and life and global climate research. In order to further promote the study of polar glaciers, we can start from the automatic processing of ice sheet radar topological sequences. Due to...

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): G06N3/04G06N3/08G06K9/62
CPCG06N3/04G06N3/08G06F18/2414G06F18/253
Inventor 蔡轶珩刘丹谢锦杨静贤
Owner BEIJING UNIV OF 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