A project to classify landmarks/natural resources in a HSI file by training a deep convolutional neural network with the extracted features.
Usage
The code is written in python. So recommended following:
- Python3
- Certain libraries - Scikit, numpy, matplolib, PIL OR just download anaconda 4 :)
- Spectral python Spy
- wxPython Download
Current implementation
-
The code for visualizing the HSI files that can generates the graphs for any pixel in an image and also the spectral signature of any class.
-
The HSI classifier code models a network to learn the features of every class's pixels ,to be able to classify any given pixel accurately. Level of accuracy - 61%
-
Tumor Image plotting is a simple workaround with the pillow library, extracting pixels from image channels and plotting a graph