Abstract
Accurate and efficient segmentation of medical serial images is of paramount importance for a wide range of clinical applications. A novel method for serial image segmentation is proposed in this paper. In the proposed framework, an adaptive external force is designed based on Laplacian of Gaussian to correctly attract the evolving curve from either side of object boundary, which makes the automatic segmentation of serial images become possible. The Tukey's bi-weight function is also introduced to match the evolving curve with object boundary, so the moving contour will stop evolution at object boundary accurately even if the blurry image is processed. Experiments on both single and serial images demonstrate the advantages of our method in terms of accuracy and robustness.
The work is supported by the National Natural Science Foundation of China (grant nos. 30801302 and 30872906) and ‘Medical and Engineering Crossing’ Foundation of Shanghai Jiao Tong University.