Significant target detection method based on multi-graph optimization and rare characteristics
A technology of target detection and characteristics, applied in the field of image processing, can solve the problems of rare characteristics, without considering the prior knowledge of visual saliency, etc., to achieve the effect of improving performance
Inactive Publication Date: 2018-05-18
SOUTHEAST UNIV
View PDF3 Cites 2 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
In recent years, salient object detection methods based on graph optimization framework have achieved certain results in simple natural scene images, but salient object detection in complex natural scene images has certain limitations.
[0004] Most of the work is optimized based on a single graph, and using a single graph to describe the information contained in an image has certain limitations
In addition, most of the work does not consider the prior knowledge of visual saliency when designing the optimization framework, such as rare features
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0036] Such as figure 1 As shown, the salient object detection method based on multi-graph optimization and rare features in this embodiment includes the following steps in sequence:
[0037] S1: Use the existing SLIC algorithm to over-segment the image into superpixels, and calculate the position and color features of each superpixel;
[0038] In this embodiment, the image to be processed is over-segmented into 200 homogeneous superpixels using the SLIC method; spatial position features and CIELab color features are extracted for the over-segmented superpixels.
[0039] S2: Define the superpixel as the node of the graph, and construct two different graphs according to the position and color characteristics of the superpixel to describe the input image;
[0040] In this embodiment, the over-segmented superpixels are defined as the nodes of the graph. Construct a graph according to the spatial position features: connect the superpixels with the adjacent superpixels to obtain ...
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
Login to View More Abstract
The invention discloses a significant target detection method based on multi-graph optimization and rare characteristics. The method comprises steps that over-segmentation of an image is carried out through utilizing an SLIC algorithm to acquire superpixels, and the position and color characteristics of each superpixel are calculated; the superpixels are defined as nodes of the graph, and two different graphs are constructed based on the positions and the color characteristics of the superpixels to describe an input image; the rare characteristics of the superpixels based on the degree of thenodes in the graph are calculated; the seed node information based on the image edge superpixels and the constructed graphs is acquired; a multi-graph optimization framework is proposed, integrating the constructed multiple graphs, the rare characteristics of each superpixel and the seed node information, significant targets in a natural scene image are detected. The method is advantaged in that the effective visual rare characteristics of the human eyes are fully considered, improvement of significant target detection performance of the complex natural scene image is facilitated, compared with other fifteen types of significant target detection methods, through verification,the detection result is more consistent with a true value graph of a database.
Description
technical field [0001] The invention relates to the technical field of image processing, in particular to a salient target detection method based on multi-image optimization and rare features. Background technique [0002] With the rapid development of information technology, the images of natural scenes captured by digital image acquisition devices (cameras, smartphones, etc.) have shown an explosive growth. In addition, due to the convenience and quickness of image collection, the content of the collected images is complex and changeable. In the context of the above image big data, how to improve the capabilities of computer vision and image processing is a common concern of academia and industry. [0003] Visual salient object detection is a powerful data analysis tool to improve the performance and speed of computer vision and image processing by extracting salient regions containing key information from images. In recent years, salient object detection methods based o...
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
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/32
CPCG06V10/255
Inventor 张金霞魏海坤谢利萍
Owner SOUTHEAST UNIV