Gastrointestinal-Disease-Classification-through-Explainable-and-Cost-Sensitive-DNN-with-SCL

Pre-trained Models

The pre-trained models that have been used are mentioned below-

• ResNet50
• EfficientNet
• DenseNet
• Inception
• Xception

Code

The "Code" folder contains the code of this project. Specific libraries, that are mentioned in the Notebook, are needed to be installed to run the code. The "Code" folder consists of "Contrastive Learning" and "Contrastive Learning with Cost-Sensitive Approach" folder.

• "Contrastive Learning" folder contains the implementation code of above-mentioned pre-trained models with contrastive learning.
• "Contrastive Learning with Cost-Sensitive Approach" folder comprises the code of cost-sensitive pre-trained Xception model with supervised contrastive learning.

Data

The employed Hyper Kvasir Dataset, which is from kaggle, can be found in the "Data" folder.