ملخص الدراسة:

In this research, new ideas are proposed to enhance content-based image retrieval applications by representing colored images in terms of its colors and angles as a histogram describing the number of pixels with particular color located in specific angle, then similarity is measured between the two represented histograms. The color quantization technique is a crucial stage in the CBIR system process, we made comparisons between the uniform and the non-uniform color quantization techniques, then according to our results we used the non-uniform technique which showed higher efficiency. To find the similarity between two signatures we made comparisons between the Euclidean distance and a new proposed similarity measurement, our proposed similarity showed better results and higher precision ratios. Also specific techniques has been used for making the represented histogram rotation-tolerant. The results showed promising performance compared to many other techniques. Every image in the database will be represented as a vector of 296 values, which makes the used space for representing images not significant and able to be compressed using traditional techniques of compression. The speed of converting the whole database does not matter because it will be done in the background. In our tests we used the Corel-1000 images database in addition to a Matlab code, and we made comparisons with other approaches like Fuzzy Club, IRM, Geometric Histogram, Signature Based CBIR and Modified ERBIR, our proposed technique showed high retrieving precision ratios compared to the other techniques.

توثيق المرجعي (APA)