Content Based Image Retrieval Using Wavelet Moments and Local Binary Patterns in CIE-L*a*b* Color Space

This paper concerns content-based image retrieval (CBIR) systems, where still images can be retrieved by similarity searching on visual features like color, texture, shape and spatial information at pixels level. In this work CBIR is accomplished by a simple method which combines local binary patterns (LBP) and wavelet moments computed in CIE-L*a*b* color space. Accuracy of the combined approach is assessed using precision and recall measurements. Experimental results on Wang-1000 database show that combined features approach presents a satisfactory progress relative to single feature based on LBP or wavelet moments regarding to precision and recall and it outperforms some of the existing approaches.