I.: Detecting edges in colour images using dichromatic differences

Many computer vision techniques need the detection of edges in the image as a first step of an image understanding process. In some cases, a grey-level edge detector is not rich enough to describe changes that are present in the image and the use of colour information and colour edges is a valuable tool for some problems. In this paper, we propose a distance function in the colour space to measure the perceived difference between two colour vectors. This function is further used combined with some concepts of the Canny edge detector to define a novel approach to colour edge detection. The experiments carried out show that the proposed method for colour edge detection provides satisfactory results both in synthetic and real colour images