Subhajit Sanyal
Current Address:
Department of Computer Science and Engineering,
Indian Institute of Technology Delhi,
Hauz Khas,
New Delhi - 110016,
India
|
subhajit@cse.iitd.ernet.in
http://www.cse.iitd.ernet.in/~subhajit/
|
|
| Research Interests |
Computer Vision and Computer Graphics with focus on Image Based Modeling and Rendering,
Walkthroughs, Geometric Vision, Active Vision, Image Processing
| Education |
Indian Institute of Technology Delhi, New Delhi, India.
Ph.D. (Near Completion)
Dissertation Area: Image Based Modeling and Rendering
Advisors: Subhashis Banerjee and Prem Kalra
Indian Institute of Technology Kanpur, Kanpur, India.
M.Sc. Physics, May 2000
Scottish Church College, University of Calcutta, Kolkata, India.
B.Sc. Physics, May 1998
| Academic Achievements |
Secured 98.32 percentile, in GATE 2000
Qualified for the UGC-CSIR National Eligibility Test (NET) for Junior Research Fellowship (JRF) 2000
Selected for National Scholarship for B.Sc. Results.
| Publications |
On Learning Shapes from Shades
Subhajit Sanyal, Mayank Bansal, Subhashis Banerjee, and Prem K Kalra.
Indian Conference on Computer Vision, Graphics and Image Processing, 2004.
Modeling of Free-Form Surfaces and Shape from Shading.
Subhajit Sanyal, Mayank Bansal, Subhashis Banerjee, and Prem K Kalra.
2nd International Symposium on 3D Data Processing, Visualization, and Transmission 3DPVT'04
To access a detailed report and see some demos, please visit.
http://www.cse.iitd.ernet.in/vglab/
Multilevel modelling and rendering of architectural scenes.
Akash M Kushal,Gaurav Chanda,Kanishka Srivastava,Mohit Gupta,Subhajit Sanyal,T. V. N. Sriram,Prem Kalra,Subhashis Banerjee. EUROGRAPHICS 2003.
To access a detailed report and see some demos, please visit.
http://www.cse.iitd.ernet.in/vglab/
| Current Research Activity |
Gaze control and automated navigation of avtars in guided tours of virtual worlds.
To be communicated shortly.
| Work Experience |
Currently employed as a Project Associate in the project ``Design and Development of Image Based Rendering System'', in the Department of Computer Science and Engineering, Indian Institute of Technology Delhi.
| Projects |
A selected list of projects with which I was associated.
Virtual Advertisements:
The project was to artificially render dynamic advertisements on some static billboards in a real scene as it is
being captured in a video-footage. The issues involved were real-time segmentation of occluders
and simple geometric etchniques to enable the transfer of the required texture on the billboard. For a detailed
report and some demos visit
http://www.cse.iitd.ernet.in/vglab/
Augmented Reality:
The project involved developing a novel technique for simple and fast augmentation of a static scene
with artificial moving objects or objects extracted from a motion clip. The technique essentially deals
with the recovery of camera parameters of the static image followed by a sparse modeling of the potential
occluders. A detailed report and some demos can be accessed at
http://www.cse.iitd.ernet.in/vglab/
Robot Navigation:
The project involved developing a robot navigation system, in which a robot, with mounted cameras,
can detect obstructions, and traverse a path avoiding them, from its starting position to the designated destination.
Some reports can be accessed at
http://www.cse.iitd.ernet.in/vglab/
Image based Rendering of Walk-through:
Recently, the problem of image based rendering has attracted considerable attention wherein the environmental
map for rendering of novel views are maintained in terms of a set of images instead of explicit geometric
and photommetric models of a scene. The project addressed the problem of generation of a sequence of novel views
of a scene from a set of reference views. To access a detailed report and some demos visit
http://www.cse.iitd.ernet.in/vglab/
Volume Rendering on GPU:
Volume Rendering has long been used for medical imaging, fluid dynamics and other such applications where information from 3D data is to be extracted. Due to large data sizes and intensive computations it has long been a classical problem in graphics. The recent emergence of hardware accelerated algorithms have furthered these applications and allows real time rendering. The project was to implement and analyse different techniques for rendering volumetric data and utilisation of GPU for achieving real-time interactive visualization.
| Computer Skills |
Platforms:
Linux, Windows
Languages:
C,C++,Java, x86 Assembly
Computer Vision and Graphics specific skills:
Cameras, frame-grabbers, OpenGL
| References |
Available on request.
|
|