A Histogram with Perceptually Smooth Color Transition for Image Retrieval

We propose a novel histogram generation technique using the HSV color space. The histogram retains a perceptually smooth color transition that enables us to do a window-based comparison of feature vectors for the purpose of effective retrieval of similar images from very large databases. During retrieval, we use a vector cosine distance measure for the ordering of image feature vectors. This distance measure shows an improvement in the retrieval result over the traditional Euclidean distance.