The technological development that the humanity has witnessed in the last few decades had opened new direction for revolutionizing healthcare. In this study, the convolutional neural network (CNN) is applied to classify and investigate the Melanoma skin cancer which is globally accounting for at least 40 % of all skin cancer cases. Through the designed Melanoma Skin Cancer Detection System (MSCDS), the classification of the inputted images into Benign or malignant tumors can be done. The MSCDS is empowered with flexible and friendly Graphical User interface (GUI) to assess the patients in the process of frequent skin cancer check-up. A CNN architecture has been developed to achieve a high level of robustness and accuracy. The accuracy of MSCDS has been compared with all the similar attempts in the literature. Furthermore, MSCDS has shown a superior accuracy of 90.83% over the other contributions that have previously been done in which are based on image processing techniques and few other machine learning algorithms. Moreover, we believe the MSCDS can be applied to related medical imaging applications.
You may refer to the Jupyter notebook file for the main code and the rest is for the GUi which was built by the help of PYcharm and PyQt 5 tools..
The picture below show some example of the benign and malignant tumors:
All the dataset has been taken from ISIC https://isic-archive.com/