Fast active contour for object tracking in image sequence

Abstract

This paper proposes a segmentation method of 2D image sequence based on active contour or snake. We define energy functional of active contour for detecting object in first frame of sequence. Then, the final snake is used as initial snake in the next frame. Our idea of tracking object is achieved by change of energy functional. Therefore, we substitute balloon energy of first energy functional by new energy. We propose a new formula of this new energy that guarantees fast tracking of movement of snake points. Hence, rapidity of our active contour is based on three notions: proposed initialization, change of energy functional and new motion energy. The efficiency of our approach is shown through some experimental results using three image sequences: synthetic sequence, moving biological cell sequence and echocardiograph image sequence. Its rapidity is justified by runtime comparison with a conventional method.