Yoji Arai1 and Ryuji Takaki2
1Kyoritsu Women's Junior College, Chiyoda, Tokyo 101, Japan
2Tokyo University of Agriculture and Technology, Fuchu, Tokyo 183, Japan
(Received September 5, 1994; Accepted March 26, 1995)
Keywords: Keywords: Connectivity, Gray Pattern, Integral Method, Image Analysis
Abstract. Abstract. A new method developed by the present authors to estimate characters of gray patterns, such as the number of separate elements and their connectivity, was applied to some examples, i.e. an artificial isolated spot with random noise, a photograph of thermal convection and cross sections of pathological human liver. In order to eliminate noise and increase contrast, image data were smoothed by averaging or taking median and gray levels were modified by the use of a mapping function. This method proved to be considerably effective for automatic characterization of gray patterns if an enough care is paid to remove noises by iterative smoothing.