Multiscale Reconstruction and Gradient Algorithm for X-Ray Image Segmentation Using Watershed

Because x-ray image has low contrast resolution and edge sharpness, this paper presents an improved marker- controlled watershed algorithm based on multiscale reconstruction and gradient. For overcoming over-segmentation, the raw image is filtered by gray reconstruction and extracted gradient with multiscale structure elements. The way of reconstruction filtering includes toggle mapping, tophat transformation and bottomhat transformation. This method not only enhances the contrast and sharpens the edges of raw x-ray image, but also preserves contours of raw image furthest. Applying multiscale gradient algorithm, which combines the advantages of small and big structure elements, could extract the better gradient image from filtered image. The proposed algorithm has been executed on the orthopedic x-ray image, the analysis of compare between normal morphological filtering and reconstructed morphological filtering is given. The experimental result shows that the marker-controlled watershed algorithm is more efficient for x-ray image by reconstruction filtering and extracted multiscale gradient. Using the proposed method can obtain the significant segmentation result.