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

Binocular salient object detection method based on boundary-aware neural network

A neural network and object detection technology, applied in image analysis, image enhancement, instruments, etc., can solve problems such as insufficient information, achieve the effect of improving accuracy, increasing precision, and increasing operating speed

Active Publication Date: 2021-11-19
ZHEJIANG UNIVERSITY OF SCIENCE AND TECHNOLOGY
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Existing algorithms only for planar images show the defect of insufficient information

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
  • Binocular salient object detection method based on boundary-aware neural network
  • Binocular salient object detection method based on boundary-aware neural network
  • Binocular salient object detection method based on boundary-aware neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

[0035] A binocular visual salient object detection technology based on a dual-stream fusion neural network proposed by the present invention includes two processes of a training phase and a testing phase, and the specific steps of the training phase process are:

[0036] Step 1: Select a database with color maps, disparity maps, and segmentation label maps; then scale all images in the database to a size of 256×256 by bilinear interpolation; then randomly take out 80% of them The color map and its corresponding disparity map and segmentation label data are used as the training set, and the kth color map in the training set is marked as Record the disparity map corresponding to it in the training set as Its corresponding segmentation label image is denoted as Among them, k is a positive integer, 1≤k≤K, K represents the total number of colo...

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 binocular salient object detection method based on a boundary perception neural network. A multi-scale convolutional neural network is used to construct a dual-stream neural network through two inputs. And in the pooling layer, the image is gradually reduced and encoded, so as to improve the receptive field of the convolution kernel. While improving the running speed of the program, it can quickly determine the position information of the object. Secondly, upsampling is used for decoding, and the pixels of the image in the detailed position are gradually restored, thereby improving the final accuracy. The neural network is optimized using disparity maps. It makes up for the shortcomings of insufficient information of pure 2D color images in the case of similar background and object color, contrast, etc. This makes the result closer to the real situation of human observing objects.

Description

technical field [0001] The invention relates to a stereoscopic image processing technology based on binocular vision, in particular to a binocular vision salient object detection method based on a boundary perception neural network. Background technique [0002] The human visual system will quickly focus on objects with certain characteristics before judging the received visual information. Prioritize it instead of processing all received messages. This has a great effect on improving the processing speed of the system and increasing the accuracy of judging objects. Extracting key areas and objects in an image, lightening and removing the background can greatly help the application of the image in various fields. [0003] However, a single analysis and calculation of the planar image cannot achieve this goal very well. Especially in scenes where the color, contrast and other indicators of the object and the background are very similar. Existing algorithms that only focus...

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): G06T7/00G06T7/155
CPCG06T7/0002G06T2207/10024G06T2207/20081G06T2207/20084G06T7/155
Inventor 周武杰陈昱臻雷景生李颜娥王海江何成
Owner ZHEJIANG UNIVERSITY OF SCIENCE AND TECHNOLOGY