Decoupled Active Contour (DAC) Optimization Using Wavelet Edge Detection and Curvature Based Resampling

Locating an accurate desired objectboundary using active contours and deformable modelsplays an important role in computer vision, particularlyin medical imaging applications. Powerful segmentationmethods have been introduced to address limitationsassociated with initialization and poor convergence toboundary concavities. This paper proposes a method toimprove one of the strongest and recent segmentationmethods, called decoupled active contour (DAC).Herewe apply Wavelet edge detection on the image whichcause it to have more contrast to have more informationabout edges, followed by an optimum updating on themeasurements using Hidden Markov Model (HMM) andthe Viterbi search as an efficient solver. In order to havea more accurate boundary at each iteration more pointsare injected in the high curvature parts based on thesnake curvature so we will have more precision in thesepart and also flat parts.

Keywords:

active contour , wavelet edge detector curvature , DAC