Histogram-based CT brain lesions image segmentation using active contours

Image segmentation plays a key role in the extraction of lesions such as tumors and haematoma from brain computed tomography images. This paper deals with this issue and proposes a method to achieve this goal by using an active contour driven by the image gradient and a smoothed histogram-based gray-values statistics. Moreover, the processing is limited to a specifical part in the image to lighten the computation load. Results seem to be very promising ones.