Siddharth Somasundaram

Media Lab grad student @sidartsoma joins the @Nature podcast to discuss new research from the @cameraculture group that demonstrates how consumer-grade LiDAR can be used for non-line-of-sight (NLOS) imaging — letting users see around corners. youtube.com/watch?v=yrczto…

@tedlasai Thanks Sai!! :)

Some crazy work from the big man Sid :)

@CartSiff Thanks carter!!

@sidartsoma Great work, congrats Sid!

I’m excited to share our Nature paper on seeing around corners with consumer LiDARs! We show that consumer LiDAR sensors — in your smartphones, AR headsets, self-driving cars, and robots — can be used to see objects hidden around corners. youtube.com/watch?v=N3LEhh…

@anagh_malik Thanks anagh!! And congrats on your sick CVPR paper

@sidartsoma looks amazing, congrats Sid!!

(n/n) Joint work with Aaron Young, @adaveiitm, @AdithyaPadire, and @raskarmit at @mit @medialab @cameraculture @dartmouth We have lots of exciting work coming out! Please reach out if you want to work together

(8/n) Check out the links below! 𝗣𝗿𝗼𝗷𝗲𝗰𝘁 𝗽𝗮𝗴𝗲: cornar.media.mit.edu 𝗖𝗼𝗱𝗲: github.com/sidsoma/consum… 𝗣𝗿𝗲𝘀𝘀 𝗿𝗲𝗹𝗲𝗮𝘀𝗲: media.mit.edu/posts/mit-medi…
Paper: nature.com/articles/s4158… 𝗮𝗿𝗫𝗶𝘃: arxiv.org/abs/2605.17865 𝗣𝗼𝗱𝗰𝗮𝘀𝘁: youtube.com/watch?v=yrczto…

In @Nature, a team of Media Lab researchers demonstrates that consumer-grade LiDAR — the same sensors that help iPhones autofocus, power depth on the Vision Pro, and guide robot vacuum cleaners across the living room floor — can also make it possible to see around corners. Using off-the-shelf hardware costing less than $100 and no specialized calibration, the team reconstructs hidden 3D objects, tracks moving targets (including a user's hands), and even uses hidden objects as landmarks to figure out where a camera is in space. This technical breakthrough in non-line-of-sight (NLOS) imaging is a new technique the team calls motion-induced aperture sampling, which turns the natural shake of a handheld device into an asset, stitching together faint signals from light that bounces off nearby walls and floors. "The most exciting part of this work to me is that we took a capability that used to require a specialized $50,000 imaging setup and put it into the hands of people in robotics, AR/VR, and beyond," says lead researcher Siddharth Somasundaram (@sidartsoma), a PhD student in the Camera Culture research group (@cameraculture). “As LiDARs become more common, I think this could lead to entirely new forms of machine vision and spatial perception.” Read more: media.mit.edu/posts/mit-medi…

📢📢📢Dark3R (#CVPR2026) recovers camera poses and point clouds—and enables novel view synthesis—from images captured in ultra-low light (image SNR < −4 dB). Project Page: andrewguo.com/pub/dark3r Code/Dataset: github.com/andrewyguo/dar… Paper: arxiv.org/abs/2603.05330

Now out in Science Advances 🎉 Evolution designs eyes; RL learns the behavior. Result: Cambrian explosion of vision and intelligence (GIF ↓) Link: science.org/doi/10.1126/sc…

👁️🌋 Our new @ScienceAdvances paper: We replayed the Cambrian explosion of vision by evolving AI agents inside a physics engine to understand the principles that shape visual intelligence. We believe that this is a promising way to do AI for science and build new forms of AI by computationally mimicking biological design principles of evolution and learning. Website: eyes.mit.edu and paper links at the end of this thread 👇

Excited to share 🏀 Shoot-Bounce-3D 🏀 at SIGGRAPH Asia '25 this week in Hong Kong! As part of it, we've released the Meta Synthetic Environments Lidar Dataset - the first large-scale synthetic single-photon lidar dataset of ~100,000 scenes! Project page: shoot-bounce-3d.github.io

Everyone says they want general-purpose robots. We actually mean it — and we’ll make it weird, creative, and fun along the way 😎 Recruiting PhD students to work on Computer Vision and Robotics @umdcs for Fall 2026 in the beautiful city of Washington DC!