INHOMOGENEOUS SURFACE-DIFFUSION FOR IMAGE FILTERING

Most of the recent work on inhomogeneous diffusion in image filtering focuses on diffusing the isotope curve. We present a less familiar app roach to the development of inhomogeneous diffusion algorithms in whic h the image is regarded as a surface in three-space. The magnitude of the surface normal controls a diffusion that evolves the image surface at a speed proportional to its mean curvature. A discrete algorithm t o implement this proposed diffusion is introduced, and we show experim entally that the algorithm develops singularities that reveal, preserv e, and enhance the underlying signal. If the input signal is an isolat ed noisy edge, this leads to complete noise removal and enhancement wi thout affecting the edge locality. (C) 1997 John Wiley & Sons, Inc.