DaneshSelwal/Classification-DIAS-65Bands

DIAS Hyperspectral Classification - Model Registry

This repository contains the trained model weights for the DIAS 65-Band Hyperspectral Classification project.

Models Included

This registry contains 38 specialized .keras models, including:

1. Baseline Architectures: AlexNet_CNN, GFNet, ViT_UNet
2. Evidential Deep Learning (CREDIT): Architectures trained with evidential loss for raw uncertainty.
3. Deep Ensembles: 5-model ensembles for epistemic uncertainty quantification.
4. Multi-Head CP Models: Architectures adapted with auxiliary dense heads for multi-class conformal prediction.

Usage

These models are designed to be used in conjunction with the codebase hosted in our Project Workspace Space.

Note

Code and datasets have been decoupled from this repository to adhere to standard industry practices.