Skin cancer detection

These implementations are based on the following research papers:

• Melanoma lesion detection and segmentation using deep region based convolutional neural network and fuzzy C-means clustering

• A comparative study of deep learning architectures on melanoma detection

Repo structure

• dataset: Scripts and instructions on how to handle images

Requirements

• Python 3.9+
• Jupyter notebook
• Anaconda or Miniconda (optional but recommended)

Install required libraries

pip3 install -r requirements.txt

1. Melanoma lesion detection and segmentation using deep region based convolutional neural network and fuzzy C-means clustering.

This implementation of skin cancer detection has three main steps namely:

1. Skin Enhancement Phase.
2. Lesion localization Phase.
3. Lesion segmentation Phase.

Skin Enhancement Phase

The first step was to preprocess / refine images to remove hair and other artifacts from then input image. We did this by running two morphological closing operations (dilation, then erosion) on the input image.