Heterogeneous distributions, Minimum Cross Entropy Thresholding, Parallel processing, Skin Cancer images.


Databases and Information Systems | Graphics and Human Computer Interfaces


Melanoma is the most common dangerous type of skin cancer; however, it is preventable if it is diagnosed early. Diagnosis of Melanoma would be improved if an accurate skin image segmentation model is available. Many computer vision methods have been investigated, yet the problem of finding a consistent and robust model that extracts the best threshold value, persists. This paper suggests a novel image segmentation approach using a multilevel cross entropy thresholding algorithm based on heterogeneous distributions. The proposed strategy searches the problem space by segmenting the image into several levels, and applying for each level one of the three benchmark distributions, including Gaussian, Lognormal or Gamma, which are combined to estimate the best thresholds that optimally extract the segmented regions. The classical technique of Minimum Cross Entropy Thresholding (MCET) is considered as the objective function for the applied method. Furthermore, a parallel processing algorithm is suggested to minimize the computational time of the proposed segmentation model in order to boost its performance. The efficiency of the proposed RPSM model is evaluated based on two datasets for skin cancer images: The International Skin Imaging Collaboration (ISIC) and Planet Hunters 2 (PH2). In conclusion, the proposed RPSM model shows a significant reduced processing time and reveals better accuracy and stable results, compared to other segmentation models.

Design/methodology – The proposed model estimates two optimum threshold values that lead to extract optimally three segmented regions by combining the three benchmark statistical distributions: Gamma, Gaussian and lognormal.

Outcomes – Based on the experimental results, the suggested segmentation methodology using MCET, could be nominated as a robust, precise and extremely reliable model with high efficiency.

Novelty/utility –A novel multilevel segmentation model is developed using MCET technique and based on a combination of three statistical distributions: Gamma, Gaussian, and Lognormal. Moreover, this model is boosted by a parallelized method to reduce the processing time of the segmentation. Therefore, the suggested model should be considered as a precious mechanism in skin cancer disease detection.

Author ORCID Identifier

Nancy Zreika https://orcid.org/0000-0003-1189-1702

Ali El Zaart https://orcid.org/0000-0002-1240-3785

Abdallah El Chakik https://orcid.org/0000-0002-6859-033X





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.