Tejaswi with HTML5 Markup

Tejaswi Agarwal Graduate student @ UW-Madison


I have been working on research and development projects spanning Parallel and High Performance Computing and Computer Vision during the course of my undergraduate studies. A short abstract of some projects are below. For more detailed information you can go to the individual project pages by clicking the title or read the papers published here.


Research Projects




1. P-HGRMS: A Parallel Hypergraph Based Root Mean Square Algorithm for Image Denoising

Guide: Dr. Rajesh Kanna, Vellore Institute of Technology, India

We worked on this project, designing a Salt and Pepper Noise Removal Algorithm from Greyscale Images. The algorithm was based on Hypergraph based image modelling. The original algorithm was scaled to a parallel model and tested on NVIDIA CUDA devices for noise removal and attenuation. Results were quite impressive as compared to the serial version with 6x-18x improvement in performance. I presented this work at the 22nd International ACM Symposium on High Performance Parallel and Distributed Computing, HPDC 2013, New York, USA. It also won the BEST POSTER Award at the conference among 26 selected posters.




2. Exploiting Data Parallelism in the yConvex Hypergraph Algorithm for Image Representation using GPGPUs

Guide: Dr. Rajesh Kanna, Vellore Institute of Technology, Chennai, India

Undergraduate Research Experience - URE 004 Project at VIT University, Chennai

In this work, we proposed a parallel approach to implement the yConvex Hypergraph model by exploiting massively parallel cores of NVIDIA's Compute Unified Device Architecture (CUDA). We perform our experiments on the MODIS satellite image database by NASA, and based on our analysis, we observe that the parallel implementation outperforms its sequential counterpart by 2 to 10 times (2x-10x). This work was presented at the 27th International Conference on Supercomputing, ICS 2013, Eugene, Oregon, USA .




6. MC-RANSAC: A Pre-processing Model for RANSAC using Monte Carlo method implemented on a GPU

Guide: Dr. K. Muthunagai, Vellore Institute of Technology, Chennai, India

Probability and Statistics Course project at Vellore Institute of Technology, Chennai, India

In this project, we present a GPU based RANSAC algorithm with pre-processing by Monte Carlo method of the assumed sample set of hypothetical inliers. Based on the results, implementation using Point Cloud Library and exploiting the data intensive tasks through the NVIDIA CUDA framework, we obtain significant improvement in the performance of plane segmentation algorithm over the randomly sampled subset of hypothetical inliers. Accepted at IEEE Mathematical Modelling and Scientific Computing 2013,MMSC 2013, Mysore, India. Paper available in the Research section.