GitHub - SoubhikMajumdar/ISIC-2018-Skin-Lesion-Classification: Skin Lesion Image Classification for Applications in Melanoma Detection. Built neural network models to learn from labelled dermoscopic image data from ISIC2018 dataset, to accurately classify skin lesions into multiple lesion categories, with a particular interest in Melanoma classification accuracy and precision.

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
NN_project_vFinal_Soubhik_Katherine.ipynb		NN_project_vFinal_Soubhik_Katherine.ipynb
ReadMe.txt		ReadMe.txt

Repository files navigation

* All necessary files under this link to folder (NNproject): https://drive.google.com/drive/folders/1xpjpC6-urB89zArEaZC8duJsmZNyJQ9O?usp=sharing
* Run NN_project_vFinal_Soubhik_Katherine.ipynb cell by cell
* Do NOT run the Data Augmentation cell (it is kept in the notebook to showcase how augmentation was done)
* Upload to Google Drive, make sure directories match
* Recommend using A100/T4 GPU on Google Colab for best performance
* InceptionV3 base model utiilzed during training and fine-tuned  


#########################################################################################

About

Skin Lesion Image Classification for Applications in Melanoma Detection. Built neural network models to learn from labelled dermoscopic image data from ISIC2018 dataset, to accurately classify skin lesions into multiple lesion categories, with a particular interest in Melanoma classification accuracy and precision.

cnn-classification data-augmentation image-classification isic-challenge lesion-detection machine-learning melanoma-classification