Automated boundary extraction and visualization system for coronary plaque in IVUS image by using fuzzy inference-based method

We propose a fully automatic plaque boundary extraction system for an intravascular ultrasound (IVUS) image aiming at practical use in clinic. The IVUS image, which is commonly used for a diagnosis of acute coronary syndromes (ACS) in the field of cardiology, has coarse-grained texture due to heavy speckle noise. A medical doctor´s interpretation of the IVUS image is disturbed frequently by the heavy speckle noise. In the proposed system, the heavy speckle noise is reduced firstly by using an anisotropic diffusion filter. Secondarily, the plaque boundary is extracted by using the Takagi-Sugeno (T-S) type fuzzy inference with a weighted separability measure and some heuristic rules. Extraction of plaque boundary is achieved fully automatically. The proposed system substantially reduces the workload of medical doctors. The effectiveness of the proposed system has been verified by the experiments using the real IVUS images.