An Adaptive Thresholds Method for Gray-Level Image Segmentation Using Type-II Fuzzy Sets

Aiming at threshold uncertainty caused by fuzziness of image for image segmentation, an adaptive thresholding method for gray-level image segmentation using type-II fuzzy sets is proposed . Fuzzy index is got using type-II fuzzy sets technique, that can overcome effectively the uncertainty, histogram peaks value of an image can be find out automatically based on maximum Mahalanobis distance and minimum classification inaccuracy rate, and the number of pixel classes is used to control the shape of fuzzy index curve of the image , fuzzy membership function window-width can be also got automatically, and optimum segmentation thresholds are determined automatically according to the fuzzy index courve of the image. Experiment shows that the proposed method is effective for gray-level images.