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

RGB-D salient object detection method based on foreground and background optimization

An RGB-D, object detection technology, applied in image data processing, instrumentation, computing and other directions, can solve the problems of limited performance, RGB-D salient object detection algorithm has not yet been opened, and achieve high precision and high recall rate. Effect

Inactive Publication Date: 2016-04-20
TIANJIN UNIV
View PDF0 Cites 29 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, its performance is still limited
[0004] So far, in the papers and literatures published at home and abroad, there is no RGB-D salient object detection algorithm based on foreground background optimization.

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
  • RGB-D salient object detection method based on foreground and background optimization
  • RGB-D salient object detection method based on foreground and background optimization
  • RGB-D salient object detection method based on foreground and background optimization

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] A RGB-D salient object detection method based on foreground and background optimization, see figure 1 , the salient object detection method includes the following steps: 101: performing initial foreground modeling based on the contrast of underlying features to obtain an initial saliency map at the superpixel level; performing middle-level aggregation processing on the initial saliency map at the superpixel level to obtain a middle-level saliency map;

[0041] 102: Introduce high-level priors to the intermediate saliency map to further improve the detection effect and generate foreground probability;

[0042] 103: Calculate the boundary connectivity of the fusion depth information, and convert the boundary connectivity into background probability;

[0043] 104: Optimizing the foreground probability and the background probability to obtain an objective function;

[0044] 105: Solve the objective function, obtain the optimal saliency map, and realize the detection of sal...

Embodiment 2

[0065] Combine below figure 1 1. The specific calculation formula introduces the scheme in embodiment 1, see the following description for details:

[0066] 201: For a given input RGB-D image, segment the RGB-D image into superpixels through an over-segmentation algorithm;

[0067] This step is specifically: express each pixel in the image as a six-dimensional feature vector, [L, a, b] is the color feature of the pixel in the color space of CIElab (a color model published by the International Commission on Illumination in 1976) , where the Lab mode consists of three channels, the first channel is lightness, namely "L". The "a" channel is from red to dark green; the "b" channel is from blue to yellow. [x, y, z] is the spatial coordinate of the pixel; then define the color distance d between two pixels (i, j) c and space distance d s ; Finally, the distance metric d between pixels is obtained.

[0068] in, d ...

Embodiment 3

[0137] This method mainly uses four indicators of accuracy (Precision), recall (Recall), area under the curve (AUC) and F-measure to quantitatively measure the effect of RGB-D salient object detection.

[0138] In statistics, the ROC curve refers to the receiver operating characteristic curve (Receiver Operating characteristic Curve), which describes the performance of a binary classifier system when the threshold is changed. In order to better measure the quality of the results expressed by the ROC curve, the embodiment of the present invention also uses the area under the curve (Area Under Curve, AUC). AUC is simply the ratio of the area of ​​the lower right corner of the PR curve to the area of ​​the entire rectangular coordinate axis plane. The PR (Precision-Recall) curve is another criterion for evaluating the performance of a binary classifier. For saliency detection, the accuracy is to calculate the number of pixels in the overlapping part of the salient objects in the...

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 RGB-D salient object detection method based on foreground and background optimization. The method comprises the following steps: initial foreground modeling is performed based on low-level feature contrast, and a superpixel-level initial salient figure is obtained; a middle-level aggregation processing is performed on the superpixel-level initial salient figure, and a middle-level salient figure is obtained; a high-level prior is introduced in the middle-level salient figure to improve the detection effect, and a foreground probability is generated; edge connectivity mixing depth information is calculated, and the edge connectivity is converted into a background probability; the foreground probability and the background probability are optimized, and a objective function is obtained; the objective function is solved, a optimal salient figure is obtained, and the detection of a salient object is realized. According to the invention, a optimization framework based on foreground and background measurement and the depth information of a scene is fully utilized by the invention, a high recall rate can be obtained, and the accuracy is high; the method can accurately position the salient object in different scenes and different sizes of objects and can also obtain nearly equal salience values in the target object.

Description

technical field [0001] The invention relates to the detection field of computer vision, in particular to a RGB-D (color and depth image) salient object detection of fusion depth information. Background technique [0002] In the field of computer vision, detecting and segmenting salient objects from natural scenes is an active topic and has produced many meaningful applications. Most current salient object detection methods utilize color information as well as various prior information to achieve better results. Although scene depth plays a very important role in the human visual system, its role in salient object detection has not been deeply explored. Currently, due to the emergence of a series of depth cameras, it is more convenient to obtain scene depth information. [0003] Scene depth is a good cue for saliency detection. Compared with 2D saliency detection, only a few works have addressed the task of saliency detection based on RGB-D data. Desingh et al. [1] A met...

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): G06T7/00
CPCG06T7/0002
Inventor 周圆陈阳崔波霍树伟侯春萍
Owner TIANJIN UNIV
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