AN Adequate Method for Color Image Retrieval based on Texture
The goal of this paper is to review techniques for content-based image retrieval by using low level feature such as
texture. The reviewed techniques are Gray-level co-occurrence matrix, Color co-occurrence matrix, Local binary pattern and
Ultra local binary pattern; all these approaches are studied and compared their performance. It is found that U-LBP is
superior than above mentioned methods and the retrieval accuracy is measured in terms of Mean Square Error (MSE), Peak
Signal to Noise Ratio (PSNR), Precision and Average Precision.
Index Terms - Content-based Image Retrieval (CBIR), Color Image, Shape, Texture, LBP, ULBP, MSE, PSNR.