Skip to content

vguptai/Melanoma-Cancer-Detection-V1

Repository files navigation

Melanoma-Cancer-Detection-V1

Melanoma Cancer is a dangerous form of skin-cancer. Though, it is a rare form of skin-cancer but it is highly fatal. In this repo, we are training a Deep CNN for finding out if a lesion is cancerous or not. We use a pre-trained Inception Model to generate as a feature extractor since our dataset is relatively small.

This article explains the approach and the results in detail.

Steps to train and test the model

  1. Delete all the checkpoints from the "model" directory before training a new model from scratch.

  2. Run the following command. This will read all the images from the dataset folder and split them into training and testing set and pickle them. This is done to avoid loading the images multiple times, so skip this step if you have already done this before.

python prepareDataSetFromImages.py 
  1. Run the following command to train the models.
python inceptionTrain.py

The file config.py contains the various parameters/flags that can be set.

Releases

No releases published

Packages

No packages published

Languages