This code is part of our paper titled Knowledge Distillation approach towards Melanoma Detection submitted in the Journal of Computers in Biology and Medicine (CIMB).
Full text of accepted version avaiable at RG.
Authors: Md Shakib Khan, Kazi Nabiul Alam, Abdur Rab Dhruba, Hasib Zunair, Nabeel Mohammed
TL;DR We propose a knowledge distillation based approach towards melanoma detection. The goal is build a model with few parameters that can accurately detect melanoma and can be easily integrated without requiring much compute. It enables us to build small and performant models which can be easily deployed in clinical settings without the need for heavy computational costs.
This code requires
- Python: 3.7
- Tensorflow: 2.1.0
- Keras: 2.3.1
- OpenCV: 4.1.0
- Download the dataset from drive or Gihub Releases
- The ISIC Database.
- Load the dataset.
- Train the
Teacher&Student
Model. - Run the pretrained models.
We report the training, testing time and accuracy on the ISIC test set.
If you use this code or models in your scientific work, please cite the following paper:
@article{KHAN2022105581,
title = {Knowledge distillation approach towards melanoma detection},
author = {Md Shakib Khan and Kazi Nabiul Alam and Abdur Rab Dhruba and Hasib Zunair and Nabeel Mohammed}
journal = {Computers in Biology and Medicine},
volume = {146},
pages = {105581},
year = {2022},
issn = {0010-4825}
}