Convolutional Neural Networks for Object Recognition

This is a MATLAB implementation of a convolutional neural network to classify images from standard datasets done as a part of our course project for CS663 Digital Image Processing.

Data Sets Used :

• MNIST [Hand written digits data] as available here.
• STL-10 [10 Class Categorized Objects] as available here.
• CIFAR-10 [10 Class Categorized Objects] as available here

Note that the data set and its preprocessed variant is included in the data directory of the repository.

Steps to Train and Test the CNN

• Clone the Repository.
• Run the cnntrain.m file with input 1, 2, 3 for different datasets.
• To visualize the Weight Filters and Pool Activation Features enable the visualize option.

Team

• Meet Shah
• Yash Bhalgat