Fuzzy Representation of Edgeness in Blurred Images
Tae Yong Kim, 한준희 · Jeongbo gwahaghoe nonmunji. so'peuteuweeo mich eung'yong · 1997
Edges in an intensity image can be used in separating object from background, but it is hard to process the edges of blurred images since they are affected by various unknown degradations. These degradations make the edges fuzzy in terms of the location and existence which may depend upon the blurring factors. In order to handle such problems, we propose a fuzzy representation method of edgeness in blurred images. This new model describes the edges as truthful as possible to the intensity function by fuzzy edgeness rather than enforcing them exact or crisp as a conventional edge representation method does. The fuzzy edgeness is represented by a membership function which is defined using the distribution of intensity gradient magnitude an the symmetricity of the intensity profiles with respect to the edge location. This representation scheme can prevent the edge detection errors from propagating to a high level process. The method is also a generalization of conventional edge representation encompassing crisp as well as fuzzy cases. This scheme may be used in various image analysis tasks.