1.  Image De-fencing using Conditional Generative Adversarial Networks

     Supervisor : Dr. Pratik Chattopadhyay

   * Development of Image De-fencing framework using Conditional Generative Adversarial Networks
     (cGANs), which leads to removal of fences/barricades from an image affected by such occlusions.

   * Implemented two cGAN based network architecture, (a) Two-stage image de-fencing network, (b) Single
      stage image de-fencing network, along with edge maps in form of Canny edge map.

   *  Created our own artificial fence segmentation dataset, and fence/de-fenced image pair dataset and
      used them for comparative analysis.