top of page
Ritwik Das
Ritwik Das
I completed my Master's studies under the advisorship of Prof. Hong Qin from the Dept. of Computer Science at Stony Brook University. My interests primarily lie in Computer Graphics and Vision and I try to keep an open mind for problems in related fields. I have worked extensively on photo-realistic rendering techniques as a student as well as in the industry. My thesis was on material acquision from scanned 3D scenes. I have also worked on other projects with Prof. Joseph S.B. Mitchell and Prof. Dimitris Samaras. Details are available in my Research Blog.
I have previously worked (2012 - 2013) as a software engineer at Microsoft R&D India and I recently spent my summer (2014) at DreamWorks Animation SKG, Glendale as an R&D Render Intern. Although I have had the opportunity to dwell in different places across India as well as in Europe, I enjoy spending my time in Kolkata (the City of Joy) which has also been my home for the longest time.
My Latest Research
REALight: A Material Acquisition and Editing System for Kinect Scenes, (Thesis)
Advisor - Hong Qin
Unlike a computer-generated world as in a game or an animated world, we live in a world with real objects which have measurable physical properties associated with them. With this in mind, we present here an acquisition system which can capture real scenes and their physical properties like geometry, color and material properties and use them to present the scene to the user in a way which is more useful than the raw scene itself. This project as the title suggests deals with the synthetic manipulation of real scenes captured from camera (specifically an RGBD camera). Since the information available from a regular color camera is hard to use for material acquisition to a high degree of accuracy, we will see how we can use an RGBD camera to perform this task. (Paper under submission)
Optimal Camera Placement and Trajectory Estimation for 3D Scenes
Advisor - Joseph S.B. Mitchell
In this project we look at novel heuristics based methods for computing stationery camera placement schemes as well as camera trajectories for 3D scenes. Some of the major applications of this include model verification, 3D scanning, security etc. This problem is very similar to the Art Gallery Problem which is famous in the field of Computational Geometry. We also address several real world variants of this problem like limited visibility guarding, partial and resource constrained coverage, budgeted camera traversal etc. Our proposed algorithm has been extensively tested on a variety of scenes and it has shown very promising results with remarkable performance advantages as compared to some of the existing algorithms. Our approach has been the first to address a wide class of computational geometry problems related to the AGP for 2D and 3D domains. (Paper under submission)
bottom of page