Abstract: Diagnosis of melanoma (skin cancer disease) is a challenging task in medical science field due to the amount and nature of the data. Skin cancer datasets are usually comes in different format and shapes including medical images, hence, data require tremendous efforts for preprocessing before the auto-diagnostic task itself. In this work, deep learning (convolutional neural network) is used to build a computer model for predicting new cases of skin cancer. The first phase in this work is to prepare images data, this include images segmentation to find useful parts that are easier for analysis and to detect region of interest in digital images, reduce the amount of noise and image illumination, and to easily detect sharp edges (boundaries) of objects. Then, the proposed approach built a convolutional neural network model which consists of three convolution layers, three max pooling layers, and four fully connected layers. Testing the model produced promising results with accuracy of 0.74. The result encourages and motivates for future improvement and research on online diagnosing of melanoma in early stages. Therefore, a web application was built to utilize the model and provide online diagnosis of melanoma.

Mohammad Ashraf Ottom, “Convolutional Neural Network for Diagnosing Skin Cancer” International Journal of Advanced Computer Science and Applications(IJACSA), 10(7), 2019. http://dx.doi.org/10.14569/IJACSA.2019.0100746

@article{Ottom2019,
title = {Convolutional Neural Network for Diagnosing Skin Cancer},
journal = {International Journal of Advanced Computer Science and Applications}
doi = {10.14569/IJACSA.2019.0100746},
url = {http://dx.doi.org/10.14569/IJACSA.2019.0100746},
year = {2019},
publisher = {The Science and Information Organization},
volume = {10},
number = {7},
author = {Mohammad Ashraf Ottom},
}