Statistical modeling for the characterization of atheromatous plaque in Intravascular Ultrasound images

The chief evolution of medical imaging gains great assistant for accurate and efficient medical diagnosis over a short period of time. The medical image analysis and processing became a lively energetic research field. Since manual processes are tedious for large data there is a need for automatic processing that helps the general public. Many medical images do not exhibit regions of uniform and smooth intensities but random structures and patterns, it is necessary to use statistical techniques to analyze and process it. The main objective of atheromatous plaque characterization is to identify fibrotic, lipidic and calcified tissue in Intravascular Ultrasound images (IVUS) automatically. In this paper, the local characterization of atheromatous plaque is proposed using the combined features of shape and scale parameters in Nakagami Distribution and the mean and standard deviation of detail subbands in Discrete Wavelet Transform (coiflet) which is the promising technique. The extracted features were given as input to the classifier using Bayesian Model. The proposed algorithm could significantly contribute to a study of plaque characterization, and consequently to an objective identification of vulnerable plaques with better accuracy.