Siddharth Somasundaram

IIT@MIT, Massachusetts Institute of Technology

Papers

Total Citations

H-Index

About

Siddharth Somasundaram is pioneering the next generation of computational imaging systems that break traditional trade-offs in speed, light efficiency, and 3D reconstruction. His research spans event-based vision, single-photon avalanche diodes (SPADs), and novel LiDAR-RGB fusion techniques. In his highly cited work on event cameras meeting SPADs (2025, 4 citations), Somasundaram addresses the fundamental conflict between low-light performance and high-speed imaging—where conventional cameras suffer from either motion blur or Poisson noise—by developing a hybrid sensor framework that achieves high-speed, low-bandwidth capture even in photon-starved conditions. His second major contribution, "Blurred LiDAR for Sharper 3D" (2025, 2 citations), introduces a diffuse LiDAR approach combined with RGB data for robust handheld 3D scanning. This technique overcomes the failure modes of RGB-only reconstruction in low-texture, low-light, or low-albedo scenes, enabling reliable surface reconstruction for virtual reality, robotics, and mobile scanning. Somasundaram’s work is already shaping how next-generation mobile devices and autonomous systems perceive the world, merging hardware innovation with intelligent algorithms to push the boundaries of real-time, all-weather 3D sensing.