Oversegmentation reduction in watershed-based grey-level image segmentation

Grey-level image segmentation is concerned with the identification of the objects of interest in a digital image. We suggest a segmentation technique based on the use of the watershed transformation to partition the image into homogeneous regions, and on the successive assignment of the regions to either the foreground or the background. Two alternative criteria to reduce the oversegmentation typically affecting the watershed transform are discussed. Region classification based on the grey-level changes of adjacent regions is presented, which works well for images where the boundary between foreground and background is perceived in correspondence with strong grey-level changes.