Multi-label Image Segmentation for Medical Applications Based on Graph-Theoretic Electrical Potentials

摘要: A novel method is proposed for performing multi-label, semi-automated image segmentation. Given a small number of pixels with user-defined labels, one can analytically (and quickly) determine the probability that random walker starting at each unlabeled pixel will first reach pre-labeled pixels. By assigning to label which greatest calculated, high-quality segmentation may be obtained. Theoretical properties this algorithm are developed along corresponding connections discrete potential theory and electrical circuits. This formulated in space (i.e., on graph) using combinatorial analogues standard operators principles from continuous theory, allowing it applied arbitrary dimension.