Local Colour Features for Image Retrieval

In image retrieval scenarios, many methods use interest point detec-tion at an early stage to find regions in which descriptors are cal-culated. Finding salient locations in image data is crucial for these tasks. Observing that most current methods use only the luminance information of the images, we investigate the use of colour in-formation in interest point detection. Based on the Harris corner de-tector, a way to use multi-channel images is explored and different colour spaces are evaluated. To determine the characteristic scale of an interest point, a new colour scale selection method is presented. We show that using colour information and boosting salient colours results in improved performance in retrieval tasks. Large scale image retrieval experiments are carried out to show the gain in retrieval rates under varying circumstances.