A Gradient Weighted Thresholding Method for Image Segmentation

Otsu method is widely used for image thresholding, which only considers the gray level information of the pixels. Otsu method can provide satisfactory result for thresholding an image with a histogram of clear bimodal distribution. This method, however, fails if the variance or the class probability of the object is much smaller than that of the background. In order to introduce more information of the image, a gradient weighted threholding method is presented, which weighs the objective function of the Otsu method with the gray level and gradient mapping (GGM) function. It makes the between-class variance of the thresholded image maximize and the threshold locate as close to the boundary of the object and the background as possible. The experimental results on optical images as well as infrared images show the effectiveness of the proposed method.