Category integrated feature and integrated category matching-based image retrieval method
A technology of comprehensive feature and image retrieval, applied in character and pattern recognition, special data processing applications, instruments, etc., can solve problems such as increasing user burden, impact of retrieval results, loss, etc.
Active Publication Date: 2016-12-07
西交思创智能科技研究院(西安)有限公司
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0007] In RBIR, manual selection of matching areas will increase the user's burden; automatic matching area selection based on regional location and area characteristics is easy to miss key areas and cause information loss; the complete area matching method has the disadvantage of repeated area matching, especially when repeated matching When the area is the background, it will have a huge impact on the search results
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
[0090] In the present invention, the CaBIR retrieval framework is proposed, the class in the image is extracted by the ASRM-AP method, and the feature extraction of the class is carried out by using IFOC to reduce the difference between low-level features and high-level semantics. The importance in assign weights to them and match them. By aggregating all regions in the image into several unique classes that can fully represent the image, extracting features at the class level and performing matching, it avoids repeated matching while ensuring the integrity of information, thereby improving the retrieval quality. The specific process is in figure 1 given in.
[0091] The steps of the present invention are as follows:
[0092] 1. Extract the class in the image
[0093] A class is a collection of similar regions in an image. In the present invention, a method combining segmentation and clustering is used to extract the class in an image. First, the image is segmented by 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
Login to View More Abstract
The invention discloses a category integrated feature and integrated category matching-based image retrieval method. The method is a category-based image retrieval method proposed for the problems of information loss easily caused by an individual region matching method and large calculation amount and similar region repeated matching of an individual region matching method in region-based image retrieval. All regions in an image are gathered into a plurality of categories capable of completely representing the image and having uniqueness, and features are extracted from levels of the categories and matching is performed; the categories in the image are obtained through an accelerated statistical region merging and affinity propagation method; the categories are subjected to feature extraction by utilizing an integrated feature method of the categories; and an integrated category matching method is used for allocating weights to the categories according to the importance of the categories in the image and performing matching. Corel-1000 and Caltech-256 image libraries are tested, and a result shows that the method has a better retrieval effect in comparison with a conventional region-based retrieval method.
Description
technical field [0001] The invention belongs to the technical field of image retrieval, and in particular relates to an image retrieval method based on class-based comprehensive features and complete class matching. Background technique [0002] With the development of computer network technology, more and more information is carried by digital images. How to retrieve images of interest to people from a large number of image databases has become a hot research field. Content-based image retrieval (Content- Based Image Retrieval (CBIR) technology has attracted much attention due to its good retrieval effect. Although CBIR technology has great advantages in many aspects compared with text-based image retrieval (Text-based Image Retrieval, TBIR) technology, it also faces some problems, the most important is the difference between low-level visual features and high-level semantics. The most common method to solve this problem is Region-based Image Retrieval (RBIR) technology. ...
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): G06F17/30G06K9/62
CPCG06F16/583G06F18/23G06F18/24
Inventor 孟繁杰单大龙石瑞霞曾萍萍王彦龙
Owner 西交思创智能科技研究院(西安)有限公司