Materials Classification using Bayesian Netowrks

A MATLAB implementation of Bayesian networks for Material Classification

Dataset

The Flickr Materials Database (FMD) was used for solving the classification problem using Bayesian networks.
The database can be downloaded from https://people.csail.mit.edu/celiu/CVPR2010/FMD/

Pre-requisites

The code was implemented on MATLAB 2021a. The following MATLAB packages were used.

Image Processing Toolbox

Computer Vision Processing Toolbox

Deep Learning Toolbox

Bayesian Network Toolbox

Parallel Computing Toolbox

Usage

1. Ensure the dataset and the downloaded Bayes network toolbox are present in the working directory
2. To perform classification using Bayesian Networks, execute the MATLAB live script textureBN.mlx within MATLAB.
3. To perform classification using neural networks, execute the MATLAB live script textureCNN.mlx within MATLAB.
4. The confusion chart and accuracy numbers for the above chosen approach will be presented.