Leaking detection for Medical image segmentation

Philippine Information Technology Journal

Lim, Joo Hwee¹  Li, Huiqi¹  Liu, Jiang¹  Damon, Wong Wing Kee¹  Racoceanu, Daniel¹ 

关键词: LMS;    Leaking detection;    Level set;    Curvature scalar;    MRI;    Segmentation;   

DOI  :  10.3860/pitj.v2i1.2563

学科分类：计算机科学（综合）

来源: Philippine Society of Information Technology Educators

【 摘 要 】

Medical image segmentation is a very important process for organ transplant and analysis. Common problems in medical image segmentation are over-segmentation and under-segmentation (leaking). LMS (Leaking detection for Medical image Segmentation) system, an approach for Medical Image segmentation with leaking detection is proposed to tackle the above problems. Possible image leaking is detected in this paper through adjusting the curvature scalar parameter in level-set based segmentation approach. To evaluate the effectiveness of the new methodology in kidney segmentation in real world environment, Magnetic Resonance Images from a local hospital are used to test the robustness of LMS. Statistically, LMS greatly reduces the under-segmentation (leaking), which maintains the over- segmentation performance compared with the conventional level set algorithm and achieves about 17% better performance than the conventional level set algorithm.

【 授权许可】

Unknown