Segmentation and Splitting of Touching Vaginal Bacteria Based on Superpixel and Effective Distance

Abstract

In this paper, a new method for segmentation and splitting of touching vaginal bacteria based on super pixel method is proposed. Feature fusion is integrated with kernel-based support vector machine (SVM) for bacteria segmentation. After segmentation by super pixel, the touching bacteria regions are further separated according to the defined effective distance. Finally, the separated bacteria are counted finally for the performance evaluation. Our experimental results show that the proposed method has achieved promising segmentation result. Moreover, compared to the state-of-the-arts method, better segmentation results have also been achieved.