Mehul Arora

I am currently a Masters student in Computer Science at University of California, San Diego, USA . Previously, I was a Computer Vision & Deep Learning Engineer at Preimage Pvt. Ltd . My work is primarily focused on developing structure-from-motion pipeline using drone images and GPS data. Previously, I was as a data scientist at the PayU.

I did my bachelor's thesis under the guidance of Prof. Dr. Cyrill Stachniss at StachnissLab, University of Bonn, Germany. My work targeted on creating a map for self-driving cars which was free of any dynamic objects. I have also worked under the guidance of Prof. Dinesh Manocha from the University of Maryland, College Park, USA on predicting the trajectories of self-driving cars. I was also selected as a Google Summer of Code recipient for The Mifos Organization in 2020.

I did my Bachelor's in Information Technology from IIIT Allahabad. I was advised by Dr. Rahul Kala and we worked on improving localization accuracy of traditional SLAM algorithm using Deep Learning.

Here is my Resume.

Email / Github / Google Scholar / Linkedin

Research

My research interests include computer vision, robotics, autonomous driving, and 3D reconstruction. I am interested in understanding the world as it is perceived by a robot.

	SLAM and Map Learning using Hybrid Semantic Graph Optimization Ambuj Agrawal, Dhruv Agarwal, Mehul Arora, Ritik Mahajan, Shivansh Beohar, Lhilo Kenye, Rahul Kala. ( equal contribution ) Mediterranean Conference on Control and Automation, 2022 Paper Improved localization accuracy of V-SLAM by injecting semantic information of detected corner points from images captured by a robot. Detected object in a scene were used for place recognition and correspondence matching which further enhanced the semantic information provided to V-SLAM module.
	Mapping the Static Parts of Dynamic Scenes from 3D LiDAR Point Clouds Exploiting Ground Segmentation Mehul Arora, Louis Wiesmann, Xieyuanli Chen, Cyrill Stachniss European Conference on Mobile Robots (ECMR), 2021 Paper / Code Static Map generation using occupancy grid mapping and ground segmentation
	Re-localization for Self-Driving Cars using Semantic Maps Lhilo Kenye, Rishitha Palugulla, Mehul Arora, Bharath Bhat, Rahul Kala, Abhijeet Nayak IEEE International Conference on Robotic Computing (IRC), 2020 Paper A novel approach of utilizing semantic information present in the environment for robot localization

Independent Projects

	See++ Code / Slides A dynamic real-time visual aid for blind people, using Object detection, image captioning and Alexa voice services.
	Credit Scoring (GSoC'20) [Project Gist] The project consisted of providing an AI powered solution to the users for credit assessment of loans.

Design and source code from Jon Barron's website