IIT Gandhinagar, Departments of CSE and EE
Computer Vision, Imaging & Graphics Laboratory
We study how images are formed, and how to invert that process to recover the shape, material and light of the scenes behind them.
- Led by
- Prof. Shanmuganathan RamanDr Vilas Mujumdar Chair, Professor and Head of CSE
- Est.
- 2013
- Record
- 140+ papers, 6 patents

Research highlights
All publicationsTensoIS: A Step Towards Feed-Forward Tensorial Inverse Subsurface Scattering for Perlin Distributed Heterogeneous Media
MERLiN: Single-Shot Material Estimation and ReLighting for Photometric Stereo
ScribGen: Generating Scribble Art Through Metaheuristics
Shadow Art Revisited: A Differentiable Rendering Based Approach
Research areas
Six threads that run the length of the imaging pipeline: capturing light, inverting it into shape and material, and rendering it back.
-

Computational Photography & HDR
High-dynamic-range imaging from ordinary cameras; tone mapping and exposure fusion.
-

Photometric Stereo & Inverse Rendering
Shape, reflectance and light from images, polarization and scattering.
-

3D, Point Clouds & Neural Rendering
Point-cloud completion, meshes and neural radiance fields.
-

Differentiable Rendering & Visual Art
Shadow art, anamorphic sculpture and scribble art as optimisation problems.
-

Saliency, Symmetry & Retargeting
Symmetry detection, saliency and content-aware retargeting.
-

Learning, Restoration & Generative Vision
Restoration, inpainting, EEG decoding and generative video.
Selected publications
All publications and BibTeX-
WACV 2027 GOT-PCD: Geometry-Guided Optimal Transport for Point Cloud Denoising
-
BMVC 2026 Why Multi-Source Audio Fails to Compose: A Geometric Diagnosis and Its Remedy
-
WACV 2026
Synthesizing Compositional Videos from Text Description
-
CVPR 2025
RASP: Revisiting 3D Anamorphic Art for Shadow-Guided Packing of Irregular Objects
-
CGF 2025
TensoIS: A Step Towards Feed-Forward Tensorial Inverse Subsurface Scattering for Perlin Distributed Heterogeneous Media
-
ECCV 2024
MERLiN: Single-Shot Material Estimation and ReLighting for Photometric Stereo
-
SIGGRAPH Asia 2024
ScribGen: Generating Scribble Art Through Metaheuristics
-
WACV 2022
Shadow Art Revisited: A Differentiable Rendering Based Approach
News
-
AwardArjun Badola and Dikshit Hegde win the Qualcomm Innovation Fellowship India 2026.
-
AwardDikshit Hegde is awarded the IndiaAI PhD Fellowship.
-
DefensePrajwal Singh defends his PhD thesis, Scalable Generative Vision and Neural Decoding.
-
DefenseSeema Kumari defends her PhD thesis on the restoration and analysis of 3D point clouds.
-
Paper acceptedLab papers accepted at IEEE ICIP 2026, ICASSP 2026, ISBI 2026 and WACV 2026.
-
DefenseAshish Tiwari defends his PhD thesis and joins Dolby Laboratories.
-
ServiceProf. Raman takes charge as Head of the Department of Computer Science & Engineering, IIT Gandhinagar.
Teaching, joining and contact
Computer Vision
Classical and learning-based vision, from image formation to generative models.
Join through IIT Gandhinagar's Summer Research Internship Programme (SRIP) or its doctoral admissions.
- Lab address
- Academic Block 13, Rooms 313 & 318, IIT Gandhinagar, Palaj, Gujarat 382355
- shanmuga@iitgn.ac.in