Image searching method based on important bit-plane

Abstract

The invention relates to a new image search method based on an important bitplane, which uses an image bit-plane theory, combines a noise attack characteristic (less effect on the bitplane), and searchs similar images by extracting content in the important bit-planes of the images. The invention mainly comprises following characteristics: at first, extracting the important bitplane from an original image; then selecting a color histogram of the important bitplane as a color characteristic, and selecting a color standard offset of the important bitplane as a space characteristic; finally, calculating similarity of contents of images by synthetically using the color and space characteristics, ordering based on the similarity, and outputting a similar image gather. The image search method of the invention not only has a simple calculation, but also has a high precision ratio and recall ratio, especially has a good robustness for noise attrack such as illumination, sharpening, fuzzy and the like.

Description

technical field [0001] The invention relates to a new image retrieval method based on important bit planes, which belongs to the interdisciplinary field of computer vision, pattern recognition, image retrieval and other fields. Background technique [0002] Since the 1990s, with the rapid development of multimedia technology and network technology (especially Internet technology), the sources of digital images are constantly expanding. In order to quickly and accurately find what users need from a large amount of image data, Content Based Image Retrieval (Content Based Image Retrieval) technology has received widespread attention and has become a hot spot in international academic research. The extraction and expression of image features is the basis of content-based image retrieval technology. The content of the image mainly includes color, texture, shape, and the spatial position relationship of sub-objects. Since color is the most intuitive feature of an image and one o...

AI Technical Summary

Problems solved by technology

To solve this problem, some methods have been proposed one after another: color histogram intersection method. Although this method has the advantages of simple feature extraction and similarity calculation, it will lose color information when performing color quantization. At the same time, the global color histogram Only the global color statistical information is recorded, but the spatial distribution information of the color is lost; the cumulative histogram method can increase the robustness of the histogram through the cumulative method, but the spatial distribution information of the color will still be lost; similar to texture symbiosis Although the matrix method can extract the color information and spatial information of the image to a certain extent, there are problems such as large computational complexity.

In addition, some researchers have proposed a method of combining color features and spatial features. Although this method considers color information and spatial information at the same time, the amount of calculation is very large. More importantly, for image libraries containing noise, accurate retrieval The rate is very low, that is, it is not robust to noisy pictures

Images