> Autonomous pathfinding in Simulated 3D Envirnment
Pathfinding in game engine environment with stereo camera input.
[Accepted at ICRAI 2017]
> Depth Map Estimation using Stereoscopic Images
Trained a convolutional neural network to predict a depth map
from given stereoscopic input images. Used batch-normalization to
speed up the process. The project was carried out under the
mentorship of Dr. OP Verma.
> Prime Minister's Office - Mobile Application
July 2015 - November 2015
One of the 5 winning submissions of a nationwide contest aimed at
developing a mobile application for the PMO. Key features were the
newsfeed, speech streaming and media galleries.
Our team was mentored by MyGov and Google Inc.
> Facial Emotion Detection Analysis
Trained different models for analysis and classification of facial
emotion into 6 categories. The Cohn Kanade image dataset was used.
The project was carried out under the mentorship of Dr. Daya Gupta.
> DTU Maps
A mobile app with to help people find places in the college campus.
Locations were crowdsourced to speed up the
creation of initial dataset.