Font Size: a A A

Based On Color Characteristics Of Image Retrieval Methods And Systems

Posted on:2010-05-24Degree:MasterType:Thesis
Country:ChinaCandidate:J ChengFull Text:PDF
GTID:2208360275955186Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the rapid development of the multimedia technology and the network technology, image data sources continue to expand,and the world's digital image capacity is growing at an alarming rate.These digital images contain a great deal of useful information,in order to find the content of user requirements from the massive database quickly and accurately,Content-based image retrieval(CBIR) technology has been widely concerned and become a hot spot of research.Content-based image retrieval makes the use of color,texture,shape and other underlying characteristics and semantic of other high-level characteristics to search image.Since the background of the image usually occupy the larger part of the image,impacts the features of the main objective of image in the process of feature extraction,images described based on global features are not well reflect the image of the semantic feature.Classical color-based histogram is easy to compute,insensitive to translation,rotation and scale,and quite resistant to noises as well. But it can't reflect spatial information of the color,prone to false hits when distinguishing images in large database with similar color composition but different spatial distribution.In order to solve this problem,after referencing to a large number of the latest research results and conduct in-depth exploration experiments,This article outlines the background and significance of image retrieval,research situation and application areas,reviews the system structure of the image retrieval,the technical classification of the retrieval method,similarity measurement methods and system performance evaluation,focuses on introducing the image retrieval base on characteristics of color,and we present an image retrieval method based on color feature and shape feature.It can use Otsu method to determine the appropriate threshold,get the background and the target image by image segmentation.As for color-based image retrieval,the paper select HSV space,at the same time the transformation between RGB space and HSV space is presented. In order to reduce the dimension of color feature,quantification of the color space to reduce the dimension.Image block firstly making the use of an improved fan segmentation strategy,and then feature extract combining color histogram,and use these features to image retrieval.This method not only can rule out the interference of background pixels,but also position the pixel of the color position of the target image.Experimental results show that this improved approach is better than the simple use of color histogram in an accurate survey.
Keywords/Search Tags:content-based image retrieval, color histogram, image segmentation, feature extraction, Eigenvector, Comparability measurement
PDF Full Text Request
Related items