About Me

I am a third year doctoral student in the School of Interactive Computing at Georgia Institute of Technology. I am advised by Prof. Jarek Rossignac. Prior to pursuing my PhD, I was a Master's student at Georgia Tech and a member of its Magic Lab and the Computer Graphics groups.

Before graduate studies, I spent three wonderful years at Microsoft India R&D Center, Hyderabad, working as a Software Engineer, and four years pursuing my undergraduate studies in Computer Engineering at Netaji Subhas Institute of Technology, Delhi.

My research focuses on developing effective 3D shape representations and accompanying algorithms for:

  • Shape statistics: Representations that facilitate computing certain statistics on collections of shapes.
  • Shape analysis: Representations that facilitate automated shape analysis.
  • Shape compression: Compressed representations for shape transmission.

I have had the pleasure of growing up in the foothills of the great Himalayan ranges, with my primary schooling years spent in the beautiful Indian hill-station of Nainital in the state of Uttarakhand. I enjoy playing table-tennis and badminton, reading the occasional book and penning down my thoughts in prose and rhyme.

Research

Shape averaging

Given two or more simple shapes what is their average shape? The answer to this question is not obvious. My recent research has been focused on providing a mathematical characterization of and fast algorithms for computing an average of a set of curves.
Relevant publications:

Mesh compression

Addressing challenges posed by the increasing complexity of 3D content requires the use of specialized data-structures for representing and transmitting triangle-meshes -- the prevalent boundary representation structures. My work proposes a random-accessible, progressive, compressed mesh transmission format. Progressive compressed mesh transmission formats address the bandwidth challenge posed in the transmission of big meshes by transmitting a compressed representation of the crude mesh initially, and, subsequentally, compressed refinements that increase the detail in the mesh. Progressive compressed formats, however, are mostly sequential, as allowing for assymetry in the decoding and encoding order increases compression costs. My work allows for the selective transmission of the desired portion of a triangle mesh. The transmitted information is compressed -- with the compression ratio matching existing progressive transmission techniques -- and also provides for random access transmission of desired regions of interest.
Relevant publications:

Hobbies

Writing

While not a prolific writer, here are three blogs that I write, in order of the frequency of their updates:

Photography

I own a treasured Nikon D5100 with two lenses - a 50mm and a 70-200mm that I have had the pleasure of taking with me to places that I travel to. My better works can be found at 500px. Here's my pick of the best of them:

Open source Development

I have made some minor contributions to a couple of community open source projects. You can get a look at these and various interesting projects that I am working on at my Github handle. I list below some direct links to a couple of interesting projects that I have worked/am working on:
  • Mesh compression and transmission
  • Raytracer
  • L-System grammar parser and tree growth simulator
  • Deep conv net for rating pictorial aesthetics