Projects
Master of Science by Research - M.S.(R) Thesis
Title: Alterations of Functional Connectivity in Autistic Brain
Master of Technology - M.Tech Thesis
Title: Static Hand Gesture Recognition Using Deep Learning
Face Tracking using Quad-rotor Drone
Programmed a quad-rotor to autonomously track the human face using its camera by changing its yaw accord-ing to the tracked face while hovering in the air. The drone turns to centre itself with the detected face. It isprogrammed to track the face of the person closest to it. The project utilised Robot Operating System (ROS),AR Drone Autonomy package and Parrot AR Drone 2.0. The project was made on Robot Operating System (ROS) using the Autonomy package. It was implemented on Parrot AR Drone quadcopter at Robotics Lab of RGUKT, R.K. Valley, Andhra Pradesh.
Depth Map Generation Using a Pair of Stereo Images
A novel parallel convolutional neural network (CNN) architecture was designed and trained on the image patches from corresponding stereo image pairs. After training, each stereo image pair was fed forward to predict a dense depth map. The results were not satisfactory and needed a higher capacity model to be trained. The work was done as part of summer research fellowship programme under the guidance of Prof. Santanu Chaudhury at IIT Delhi.
CRT: Cluster Reporting Tool
Cluster Reporting Tool is a python based tool for automatic generation of report specifying the details of the clusters of brain activations spanning multiple regions of the brain. #Neuroimaging
Bachelor of Technology - B.Tech Project
Title: Barcode Enabled Student Registration and Course Management System.
Supervisor: Prof. Rahul garg, IIT Delhi
Description: This is an investigative study that reviews the past functional MRI literature to construct a graph describing consistent alterations in the connectivity between brain regions in autistic participants. It also investigates the inconsistencies in the literature by modelling the connectivity as a function of various confounding pre-processing variables.
Github
Thesis
Supervisor: Prof. G.C. Nandi, IIIT Allahabad
Description: Movement of a differential wheeled robot was controlled by just using hand gestures. We used Microsoft Kinect mounted on movable robot named Social Mobile Autonomous Robot Test-bed (SMART) developed in Robotics and Artificial Intelligence Lab of IIIT-Allahabad for tracking the hand. We proposed a deep neural network based on the combination of denoising and sparse autoencoders for fea-ture extraction and gesture recognition.
Video
Thesis
Video
Report
Supervisor: Mr. Pankaj Tiwari
Description: It involved the automation of semester-wise student registration (using barcode scanner) and course selection in the college. The project was a Windows application based on 3-tier architecture i.e. client, server and database, developed using .NET framework using C#. Microsoft SQL Server was used as the database to store the students and course details .