ABSTRACT:- The hemorrhagic stroke is caused by the rupture of the cerebral blood vessels. The blood clot formed is called hematoma. The diagnosis of hemorrhagic stroke is the diagnosis of the hematoma. One of the essential parameters for diagnosis is the topography. The knowledge of the topography of the hematoma is crucial information to the care because it is decisive on the type of rehabilitation to adapt to the victim. In this article, we have researched the methods for efficiently detecting the contours of cerebral CT images of hemorrhagic stroke. Four methods were used, namely the gradient methods of Prewitt, Sobel, and Roberts, and Canny’s method. The results obtained have shown that the four methods allow the overall detection of contours, but that Sobel and Prewitt have thick contours and are therefore not suitable for CT scan images. Roberts has fine contours but does not accurately detect all forms of the hematoma. Only Canny’s method gives an almost perfect detection of the contours of the hematoma and is therefore the most suitable method for detecting contours of cerebral CT images.

Key words: Hemorrhagic stroke, Hematoma, topography, contour detection