A novel method for image thresholding using interval type-2 fuzzy set and Bat algorithm

In this paper, we propose a novel image thresholding method based on the interval type-2 fuzzy set (IT2FS). The interval type-2 fuzzy membership function (IT2FMF) is generated from a bag of type-1 fuzzy membership functions (T1FMFs) chosen adaptively based on the image characteristics for a given problem. An evolutionary algorithm called Bat algorithm is used to enhance the computational efficiency of the proposed thresholding technique. It is expected that the IT2FS based threshold technique will be better than that of the methods based on type-1 fuzzy set due to superior uncertainty handling capacity of the former technique. This fact is experimentally verified using benchmark dataset. The performance and robustness of the proposed method under different noise corruptions are measured statistically on the dataset by modified Cramer-Rao bound. The segmentation performance of the proposed method is compared experimentally with that of the state-of-the-art methods based on fuzzy and non-fuzzy frameworks. It is observed that the proposed method can achieve a higher segmentation accuracy in comparison to other state-of-the-art methods when they are benchmarked against the Modified Cramer-Rao Bound. Also, on average the misclassification error (ME) of the proposed method in segmentation is found to be minimum in comparison to the state-of-the-art methods. (C) 2017 Elsevier B.V. All rights reserved.