This repository contains implementation of ideas presented in the following three papers:
*Classification of Hyperspectral Images with Regularized Linear Discriminant Analysis
*HybridSN: Exploring 3D-2D CNN Feature Hierarchy for Hyperspectral Image
The performance of these algorithms are tested on publicaly available dataset of Indian Pines and Pavia University
The aim of the study consisted of two obejctives:
- Introduce novelty in the presented ideas in the first two papers to improve prediction accuracy.
- Showcase imporvement in classification accuracy by using Deep-Learning over traditional Machine Learning methods
Please refer to Report.docx for further information about implementation procedure,theoretical background and results achieved.