Nakul Agarwal

My internship is primarily focused on VO and SLAM. This includes building up a graph-SLAM solution with graph optimization using g2o. Along with that, deep learning methods are also being explored. Apart from SLAM, I was assigned other tasks which included working on deep learning (object detection, GANs etc.) and image processing.

Visited National Ilan University, Taiwan as a summer research intern. Worked on an independent project based on automating the process of noting readings from Water meters installed at houses and commercial sites. A generalized algorithm was developed which could be tweaked to work with any other kind of meters. Number recognition model was also developed by training a neural network from scratch to get an accuracy of 98.93% on testing dataset.