Medial description of gray-scale image structure by gradient-limited diffusion

摘要: The representation of object shape in grayscale images is an important precursor to many common image interpretation needs, including recognition, registration, and measurement. Typically, such computer vision tasks have required the preliminary step segmentation, often via detection edges. This paper presents a new means describing that obviates need for edge-finding, i.e., classifying pixels as being inside or outside boundary. Instead, this technique operates directly on intensity distribution produce set `medialness' measurements at every pixel across multiple spatial scales capture more global properties shape. application orientation-sensitive, gradient-limited diffusion model provides benefits global, multiscale structural analysis while preserving local region- insulating effects having edge-like properties. result procedure medial axis (MMA), which represents simultaneous levels scale, several desirable forms.© (1992) COPYRIGHT SPIE--The International Society Optical Engineering. Downloading abstract permitted personal use only.