Automated 3-D segmentation of internal hemoglobin in TEM images

Active contour models or snakes are widely used for medical image segmentation, due to their robustness to images with weak borders and poor contrast, and their ability to provide smooth contours. However, initialization is a serious problem for active contours and they tend to be attracted towards inappropriate image features if the initialization is not proper. In this paper, we propose a method that combines region growing and active contours for segmenting internal hemoglobin in transmission electron micrograph images of malaria parasites. Region growing is used in each image slice to provide the initial contour for the snake, which then determines the final contour based on the gradient vector field. Experimental results demonstrate the effectiveness of the proposed scheme in terms of segmentation quality and accuracy.