Color Image Classification and Parameter Estimation in a Markovian Framework

In this paper, we propose an unsupervised color image classification algorithm based on a Markov random field (MRF) model. In the MRF model, we use the CIE-luv color metric because it is close to human perception when computing color differences. On the other hand, intensity and chroma information is separated in this space. Without parameter estimation, our model would not be useful in real-life applications. We propose herein a new method to estimate mean vectors effectively even if the observed image is very noisy and the histogram does not have clearly distinguishable peaks. These values are then used in a more complex, iterative estimation process as initial values. The only parameter supplied by the user is the number of classes. All other parameters are estimated from the observed image. The algorithm has been tested on a variety of real images (indoor, outdoor), noisy video sequences and noisy synthetic images.