Suryansh Kumar

Résumé (updated Nov. 2023)

                suryanshkumar (at) tamu [dot] edu

I am an Assistant Professor of Visual Computing and Computational Media at Texas A&M University College Station, where I also direct Visual and Spatial AI Lab. I primarily conduct research in the field of 3D computer vision, Visual AI, and Robotic Automation. As a researcher, I am fascinated by how numerical construction can precisely represent the perceptual concepts of images, such as 3D scene geometry, motions, lights, material, and color. I aim to use these mathematical concepts to enable machines for a broader adoption using visual data. My fascination led me to explore well-developed computing fields like computer vision, artificial intelligence, computer graphics, and robotics. My research in computer vision and computer graphics aims to introduce new methods for visual representation learning, photogrammetry, and dynamic scene modeling. In AI and robotics, my research seeks to solve real-world robotic automation problems by leveraging the benefits of deep neural networks in learning visual representation and decision-making tasks.

Mobile Robotic Multi-View Photometric Stereo

Author : Suryansh Kumar
International Society Journal of Photogrammetry and Remote Sensing (ISPRS), Elsevier, 2025, IF: 10.6

Stereo Risk: A Continuous Modeling Approach to Stereo Matching

Authors : Ce Liu*, Suryansh Kumar*, Shuhang Gu, Radu Timofte, Yao Yao, Luc Van Gool. (Oral Presentation, Top 1.5% of the papers)
*Equal Contribution
International Conference on Machine Learning (ICML), PMLR, 2024, Vienna, Austria

ICGNet: A Unified Approach for Instance-Centric Grasping

Authors : René Zurbrügg, Yifan Liu, Francis Engelmann, Suryansh Kumar, Marco Hutter and others.
International Conference on Robotics and Automation (ICRA), IEEE, 2024, Yokohama, Japan

Learning Robust Multi-Scale Representation for Neural Radiance Fields from Unposed Images

Authors : Nishant Jain, Suryansh Kumar*, Luc Van Gool. (Oral Presentation, Invited)
International Journal of Computer Vision (IJCV), Springer, 2024, IF: 19.5

How To Not Train Your Dragon: Training-free Embodied Object Goal Navigation with Semantic Frontiers

Authors: Junting Chen*, Guohao Li*, Suryansh Kumar, and others (*Equal Contribution)
Robotics Science and Systems (RSS), RSS Foundation, 2023, Daegu, South Korea.

VA-DepthNet: A Variational Approach to Single Image Depth Prediction

Authors: Ce Liu, Suryansh Kumar*, Shuhang Gu, Radu Timofte, Luc Van Gool. (Spotlight Oral Presentation)
International Conference on Learning Representations (ICLR), 2023, Kigali, Rwanda.

Uncertainty Guided Policy for Active Robotic 3D Reconstruction using Neural Radiance Fields

Authors: Soomin Lee*, Le Chen*, Jiahao Wang, Alex Liniger, Suryansh Kumar, and others (*Equal Contribution)
IEEE RAL, International Conference on Robotics and Automation (ICRA), IEEE, 2023, London, UK. (Oral Presentation)

Uncertainty-Driven Dense Two-View Structure from Motion

Authors: Weirong Chen, Suryansh Kumar, and others (Oral Presentation)
IEEE RAL, International Conference on Intelligent Robots and Systems (IROS), IEEE, 2023, Detroit, USA.

Organic Priors in Non-Rigid Structure from Motion

Authors: Suryansh Kumar, Luc Van Gool. (Oral Presentation, Top 2.7% of the papers)
European Conference on Computer Vision (ECCV), 2022, Tel-Aviv, Israel.

A Real-Time Online Learning Framework for Joint 3D Reconstruction and Semantic Seg. of Indoor Scenes

Authors: Davide Menini, Suryansh Kumar*, and others. (Oral Presentation)
IEEE RAL, International Conference on Robotics and Automation, (ICRA), IEEE, 2022, Philadelphia, USA.

Jumping Manifolds: Geometry Aware Dense Non-Rigid Structure from Motion

Author: Suryansh Kumar. (Oral Presentation, Invited at Dynavis 2019)
Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, 2019, CA, USA.