Melanoma Classification Using Mixed Data in Keras

Build Keras model using mixed and imbalanced data of medical imaging and patient data to determine the presence of Melanoma.

Skin cancer is the most prevalent type of cancer. Melanoma, specifically, is responsible for 75% of skin cancer deaths, despite being the least common skin cancer. Melanoma is a deadly disease, but most cases of Melanoma can be cured with minor surgery if caught early. Dermatologists could enhance their diagnostic accuracy if detection algorithms consider “contextual” images within the same patient to determine which images represent a melanoma. If successful, classifiers would be more accurate and could better support dermatological clinic work.

The Kaggle dataset for SIIM-ISIC Melanoma Classification contains the medical images of skin lesions and patient data that we will use as inputs to determine which instances are likely to represent a Melanoma.

Here we will pass the skin lesion images to the ResNet50 model as well as pass the patient contextual data to an MLP to perform the binary classification to identify the presence of Melanoma. The Model structure is shown below.