Minimum Cross Entropy Thresholding Using Entropy-Li Based on Log-normal Distribution for Skin Cancer Images

Accurate thresholding of skin cancer images is a very essential issue in medical diagnostic applications and biometric authentication and verification systems [1], another important issue in such applications relates to its computational complexity. The objective of this study is to develop a more accurate and faster solution for estimating the optimal thresholding value of skin images. This work proposes a new algorithm for an Iterative Minimum Cross Entropy Thresholding method based on Log-normal distribution(MCET-Lognormal), under the assumption that the data of skin images is best modelled as a mixture of Log-normal distributions. The proposed method was applied on bi-modal skin cancer images and promising experimental results were obtained. Evaluation of the resulting segmented skin images shows that the proposed method yields better estimation of the optimal threshold than does the same MCET method with Gaussian distribution (MCET-Gaussian). It also reduces computational time compared to sequential search techniques.