Edge detection method insensitive to the light and shade variance in image

The article is concerned with a novel edge detection method derived from a vision model that was proposed by the authors (T. Sakamoto and T. Kato, 1998; 1999). This model is quite sensitive to abrupt changes of the image contrast and has the unique characteristic of enhancing less contrasted regions and inhibiting relatively high and firm contrasted regions on the target image. Consequently, the proposed method is insensitive to and not so much affected by illumination variance (light and shade in image) in comparison with other conventional and known methods. The paper also shows that any differential operator can be combined with this model. The authors´ proposed method can change the characteristics of target operators and enhance their abilities. These modified operators will be robust against illumination variance and detect objects in detail. The paper shows the capabilities of the proposed method by showing some examples of edge detection and making a comparison with other differential operators and their edge detection results