Finding-needles-in-a-compressed-haystack

This repo contains the course project for CS754- Advanced Image Processing.
Problem Description:
In this project I had reproduced the results of the paper " finding needles in a cmpressed haystacks" which introduced compressed learning, a linear dimensionality reduction technique for measurement-domain pattern recognition in compressed sensing applications. They showed that a large family of compressed sensing matrices satisfy the required properties. Along with the Delsarte geothal matrix, I had also implemeted bernaulli matrix with DCT basis.
Dataset used : Original Broadatz and CIFAR 10.
Algorithms used: SVM and ANN.
Note: Code is implemeted on google colab (.ipynb) notebook, so for better visualization of results open "Project_AIP.ipynb" using Google colab.