Sriram Narayanan photo

Sriram Narayanan

PhD Student, Robotics Institute, Carnegie Mellon University

I am an AI researcher working across vision, robotics, and agentic systems. I am broadly interested in building intelligent systems that can better understand the physical world, reason about people and other agents, and operate reliably in real-world environments.

Currently, I am a Ph.D. student in the Robotics Institute at Carnegie Mellon University, advised by Prof. Srinivasa Narasimhan. My research explores new ways of sensing and understanding the world, including physically grounded methods that use heat and light transport to recover scene geometry, material properties, and semantics. More recently, I have also been working on agentic AI, embodied systems, and physics-aware models for perception, interaction, and scientific discovery.

Before starting my Ph.D., I was a Research Scholar at NEC Laboratories America, where I worked with Prof. Manmohan Chandraker on trajectory prediction, embodied AI, and autonomous driving.

You can reach me at snochurn [at] cs.cmu.edu.

News

  • 06/26 Dual Band Thermal Videography won the Best Poster Award at the Computational Cameras and Displays (CCD) workshop, CVPRW 2026! 🎉
  • 04/26 Dual Band Thermal Videography accepted to CVPR 2026 as an (Oral) and named a Best Paper Candidate! 🎉
  • 03/26 PhyCo and Dual Band Thermal Videography accepted to CVPR 2026.
  • 06/25 Paper on resolving ambiguities in shape-from-heat conduction accepted to ICCP 2025.
  • 07/24 Paper Shape from Heat Conduction accepted to ECCV 2024 (Oral).
  • 02/24 Paper on visible-thermal light transport accepted to CVPR 2024.
  • 01/24 Paper on long-horizon object transport accepted to ICRA 2024.
  • 08/22 Moved to Pittsburgh and started my PhD at RI, CMU.
  • 10/21 Talk on Predicting simultaneous multi-hypotheses futures at Robotics Research Group, IIT BHU.
  • 02/21 Paper Divide and Conquer for Lane-Aware Diverse Trajectory Prediction accepted to CVPR 2021 (Oral).
  • 07/20 Paper SMART: Simultaneous Multi-Agent Recurrent Trajectory Prediction accepted to ECCV 2020.
  • 07/19 Joined NEC Labs America as a Research Scholar.
  • 06/19 Paper Talk to the Vehicle: Language Conditioned Autonomous Navigation accepted at IROS 2019.
  • 04/19 Paper A Hierarchical Network for Diverse Trajectory Proposals accepted at IV 2019.

Publications

2026

Mani Ramanagopal, Akihiko Oharazawa, Sriram Narayanan, Zeqing Yuan, Srinivasa G. Narasimhan
International Conference on Computational Photography (ICCP)
Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2026
Shows that a per-pixel temporal transform followed by spatial denoising in the coefficient domain can extract rich spatiotemporal heat flows from noisy, low-cost microbolometric videos.
PhyCo: Learning Controllable Physical Priors for Generative Motion
Sriram Narayanan, Ziyu Jiang, Srinivasa G. Narasimhan, Manmohan Chandraker
Computer Vision and Pattern Recognition (CVPR), 2026
PhyCo learns controllable physical priors—friction, restitution, deformation, and force—from simple simulations, enabling physically consistent and continuously controllable video generation without a simulator at inference.
Dual Band Thermal Videography: Separating Time-Varying Reflection and Emission Near Ambient Conditions
Sriram Narayanan, Mani Ramanagopal, Srinivasa G. Narasimhan
Computer Vision and Pattern Recognition (CVPR), 2026 🏆 Oral
🥇 Best Paper Candidate🏅 Best Poster Award · CCD, CVPRW '26
Proposes a physics-based optimization method that leverages both spectral and temporal cues from heat and light transport to effectively separate time-varying reflection and emission near ambient conditions.
A Theory of Shape Reconstruction from Heat Conduction and Shading
A Theory of Shape Reconstruction from Heat Conduction and Shading
Akihiko Oharazawa, Sriram Narayanan, Mani Ramanagopal, Srinivasa G. Narasimhan
Transactions on Pattern Analysis and Machine Intelligence (Under submission), 2026
Heat + light together resolve shape ambiguity—recovering accurate 3D geometry from a single thermal video and image without priors or calibrated lighting.

2025

Resolving Ambiguities in Heat Conduction and Shading
Akihiko Oharazawa, Sriram Narayanan, Mani Ramanagopal, Srinivasa G. Narasimhan
International Conference on Computational Photography (ICCP), 2025
Resolves local convex/concave shape ambiguities by combining shading cues with heat conduction, enabling robust 3D shape recovery without smoothness priors.
Indoor Light and Heat Estimation from a Single Panorama
Guanzhou Ji, Sriram Narayanan, Azadeh Sawyer, Srinivasa G. Narasimhan
International Symposium on Visual Computing (ISVC), 2025
Directly estimates indoor light and heat maps from HDR panoramas using image-based rendering with 3D room layout estimation.

2024

A Theory of Joint Light and Heat Transport for Lambertian Scenes
Mani Ramanagopal, Sriram Narayanan, Aswin C. Sankaranarayanan, Srinivasa G. Narasimhan
Computer Vision and Pattern Recognition (CVPR), 2024
Establishes the theoretical connection between visible light transport, thermal infrared, and heat transport in solids, proving intrinsic image decomposition is well-posed.
Shape from Heat Conduction
Sriram Narayanan, Mani Ramanagopal, Mark Sheinin, Aswin C. Sankaranarayanan, Srinivasa G. Narasimhan
European Conference on Computer Vision (ECCV), 2024 🏆 Oral
A novel shape recovery method using heat conduction from thermal videos. Estimates intrinsic shape Laplacian and recovers fine 3D surface details from simple light-bulb illumination.
Long-HOT: A Modular Hierarchical Approach for Long-Horizon Object Transport
Sriram Narayanan, Dinesh Jayaraman, Manmohan Chandraker
International Conference on Robotics and Automation (ICRA), 2024
A modular hierarchical framework for long-horizon embodied object transport, using topological scene graphs and weighted frontier exploration.

2022

Divide-and-conquer for lane-aware diverse trajectory prediction
Sriram Narayanan, Ramin Moslemi, Francesco Pittaluga, Buyu Liu, Manmohan Chandraker
US Patent App. 17/521,139, 2022
Patent for diverse lane-aware trajectory prediction using divide-and-conquer initialization.

2021

Divide-and-Conquer for Lane-Aware Diverse Trajectory Prediction
Sriram Narayanan, Ramin Moslemi, Francesco Pittaluga, Buyu Liu, Manmohan Chandraker
Computer Vision and Pattern Recognition (CVPR), 2021 🏆 Oral
A Divide-And-Conquer initialization for winner-takes-all objectives, producing diverse mode-accurate trajectory predictions anchored to lane centerlines.
Multi-agent trajectory prediction
Sriram Narayanan, Buyu Liu, Ramin Moslemi, Francesco Pittaluga, Manmohan Chandraker
US Patent App. 17/187,157, 2021
Patent for simultaneous multi-agent trajectory prediction in driving scenarios.
Simulating diverse long-term future trajectories in road scenes
Sriram Narayanan, Manmohan Chandraker
US Patent App. 17/090,399, 2021
Patent for generating diverse long-term trajectory simulations in road environments.

2020

SMART: Simultaneous Multi-Agent Recurrent Trajectory Prediction
Sriram Narayanan, Buyu Liu, Francesco Pittaluga, Manmohan Chandraker
European Conference on Computer Vision (ECCV), 2020
ConvLSTM with novel state pooling predicts scene-consistent trajectories for all agents simultaneously in constant time, with a CVAE for diverse multi-modal outputs.
Method and a system for hierarchical network based diverse trajectory proposal
Brojeshwar Bhowmick, K. Madhava Krishna, Sriram Narayanan, Gourav Kumar, Abhay Singh, M. Siva Karthik, Saket Saurav
US Patent App. 16/894,411, 2020
Patent for hierarchical network architecture generating diverse trajectory proposals.

2019

Talk to the Vehicle: Language Conditioned Autonomous Navigation of Self Driving Cars
Sriram Narayanan*, Tirth Maniar*, Jayaganesh Kalyanasundaram, Vineet Gandhi, Brojeshwar Bhowmick, K Madhava Krishna
International Conference on Intelligent Robots and Systems (IROS), 2019
Blends natural language commands with 3D semantic maps to generate local trajectories for autonomous vehicle navigation.
A Hierarchical Network for Diverse Trajectory Proposals
Sriram Narayanan, Gourav Kumar, Abhay Singh, M. Siva Karthik, Saket Saurav, Brojeshwar Bhowmick, K. Madhava Krishna
Intelligent Vehicles Symposium (IV), 2019
A two-stage CNN that maps perceived surroundings to multiple diverse trajectory proposals for robot navigation.
Gradient Aware - Shrinking Domain based Control Design for Reactive Planning Frameworks used in Autonomous Vehicles
Adarsh Modh, Siddharth Singh, A. V. S. Sai Bhargav Kumar, Sriram Narayanan, K. Madhava Krishna
Proceedings of the Advances in Robotics (AIR), 2019
A gradient-aware control law for longitudinal speed control that reactively integrates road surface gradient.

Selected Projects

Building Perception, Planning and Control for Autonomous Vehicles

Building Perception, Planning and Control for Autonomous Vehicles

2019

Bachelor thesis at IIIT Hyderabad. Designed low-level controller from scratch for a stock car, integrated perception (ORB-SLAM, LOAM) and planning (RRT*) modules, and demonstrated static obstacle avoidance in tight spaces.

Bothoven - Eyantra Robotics Competition

Bothoven - Eyantra Robotics Competition

2017

A multi-robot task execution problem where two robots simultaneously plan paths, coordinate, and play an audio file concurrently. Uses FFT to decode notes and dynamic replanning on obstacle encounters.

Swarath - Self Driving Car

Swarath - Self Driving Car

2017

Internship project at IIIT Delhi during Summer 2017, where I worked on building an SDV prototype. Designed controller and planner for the vehicle, built perception systems using sensor outputs for occupancy, and demonstrated autonomous navigation with GPS waypoints.

View All Projects →

Professional Services

Conference Reviewer:

  • CVPR: IEEE/CVF Conference on Computer Vision and Pattern Recognition (2021-Present)
  • ICCV: IEEE/CVF International Conference on Computer Vision (2021-Present)
  • ECCV: European Conference on Computer Vision (2022-Present)
  • TCI: IEEE Transactions on Computational Imaging (2025)
  • ICRA: IEEE International Conference on Robotics and Automation (2020, 2022)
  • IROS: IEEE/RSJ International Conference on Intelligent Robots and Systems (2021)
  • AAAI: Association for the Advancement of Artificial Intelligence Conference (2022, 2025)