English Title

Cross Entropy Thresholding For Skin Cancer Detection


Image Thresholding, Minimum Cross Entropy, Melanoma, Skin Images, Gamma, Gaussian, Log Normal and Poisson distributions, Bimodal technique


Architecture | Business | Engineering | Physical Sciences and Mathematics

English Abstract



Image processing procedures in medical diagnosis are used to improve diagnosis accuracy. An example of this is skin cancer detection using the thresholding approach. Thus, research studies involved in identification of inherited mutations predisposing family members to malignant melanoma have been performed in the Cancer Genetics field. Melanoma is one of the deadliest cancers, but could be cured when diagnosed early. A fundamental step in image processing is segmentation that includes thresholding, among others. Thresholding is based on finding the optimal thresholds value that partitions the image into multiple classes to be able to distinguish the objects from the background. The algorithm developed in this work is based on Minimum Cross Entropy Thresholding (MCET) method, using statistical distributions. We improved the previous work of Pal by using separately different statistical distributions (Gaussian, Lognormal and Gamma) instead of Poisson distribution. We applied our improved methods on bimodal skin cancer images and obtained promising experimental results. The resulting segmented skin cancer images, using Gamma distribution yielded better estimation of the optimal threshold than does the same MCET method with Lognormal, Gaussian and Poisson distribution.





To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.