Application of Teaching Learning Based Optimization in Multilevel Image Thresholding

This paper proposes a Teaching learning-based optimization (TLBO) algorithm for the multilevel image thresholding using Kapur entropy. In image processing, the thresholding arises to help medical imaging, detection, and recognition in making an informed decision about the image. However, they are computationally expensive reaching out to multilevel thresholding since they thoroughly search the optimal thresholds to enhance the fitness functions. In order to validate the chaotic characteristic of multilevel thresholding, a TLBO algorithm is modeled. The proposed model is an algorithm-specific, parameterless algorithm that does not require any algorithm-specific parameters to be controlled by maximizing the Kapur entropy of various classes for image thresholding. The proposed model is compared with recent algorithms to threshold the seven standard benchmark and three test images. The simulation results have higher fitness function values even with the increase of the threshold number with less computation time. The Jaccard measure values are close to 0.99.