STARS Laboratory

📅 FaVoR Presentation: Nectar Track Poster Session – Mar 22 Huge congratulations to all our authors! Looking forward to connecting with the 3D vision community in Vancouver. Drop by our sessions to say hello. 👋 🧵4/4 @UofT @UofTRobotics #EventCameras #Robotics #ComputerVision

2️⃣ FaVoR: Features via Voxel Rendering for Camera Relocalization Thrilled that FaVoR (WACV 2025) is in the 3DV Nectar Track! We explore robust camera relocalization by leveraging voxel rendering for high-quality feature generation. 🔗 papers.starslab.ca/favor/ 🧵👇3/4

🚀 Catch STARS Lab at #3DV2026 in Vancouver! We are happy to present two of our team's recent works. 1️⃣ VibES: Induced Vibration for Persistent Event-Based Sensing 🔗 papers.starslab.ca/vibes/ 🧵👇 1/4

🚨 We're recruiting MASc & PhD students for 2026! Work at the intersection of robot perception, planning & manipulation—from theory to real robots. We welcome physics, math, CS & engineering applicants. 2 weeks left to apply: utias.utoronto.ca/admissions/

We're excited to welcome Alessandro Potenza to our group! Alessandro will be studying machine learning on manifolds for robot manipulation. Welcome, Alessandro! 👋

Reminder: @trevorablett will present our work on Multimodal and Force-Matched Imitation Learning at #ICRA25 tomorrow in Vision-Based Tactile Sensors 2, Session TuBT4, room 304, at 11:30 a.m. Stick around to ask questions during the poster session!

Reminder: @trevorablett will present VPACE at #ICRA25 tomorrow in Reinforcement Learning 2, Session TuBT21, room 410, at 11:15am. Stick around to ask questions during the poster session! Pre-recorded version link below 👇

Push it to the (demonstrated) limit! To complete difficult tasks with complex contact, our latest work combines a tactile force matching algorithm with imitation learning (IL). @trevorablett will present this T-RO work at #ICRA25 tomorrow! Learn about it: papers.starslab.ca/sts-il/

Just show me what you want, and I'll figure out how to do it...quickly! @trevorablett will present Efficient Imitation without Demonstrations via Value-Penalized Auxiliary Control from Examples (VPACE) at #ICRA25 tomorrow! Check it out: papers.starslab.ca/vpace

Katie Allison will present her work at @IEEERoboSoft tomorrow in Session BB (3:15-4:15)!

Get a grip! We present a practical and inexpensive gripper fingerpad design that enables active friction tuning through structured pneumatic deflection of the fingerpad surface. MASc graduate Katie Allison will present this work at @IEEERoboSoft! Paper: arxiv.org/abs/2502.00926

We're excited to welcome Phone Thiha Kyaw to our group! Phone wil be working on long-horizon task and motion planning. Prior to joining STARS, he was with Dyson in Sinagpore. Welcome Phone! 👋

📢 FaVoR: Features via Voxel Rendering for Camera Relocalization We’re excited to share that @polivicio will present our latest work at #WACV25! 📅 March 1st, Oral Session 1.1 – 3D Computer Vision I Kudos to @marcocannici @davsca1 @domo_mr_roboto youtu.be/rgV8TZzX7qc

Congratulations to Katie Allison for succesfully defending her Master's thesis! 🥳🎓 Katie developed a structured pneumatic fingerpad for robotic grasping that enable tunable grip friction (for next-level in-hand manipulation). Each fingerpad costs less than 50 cents to make!

🎄✨ An early Christmas gift from our lab! Our paper "FaVoR: Features via Voxel Rendering for Camera Relocalization" is accepted at #WACV2025! Led by @viciopoli, with @marcocannici & @davsca1! 🎁Preprint, code and models: papers.starslab.ca/favor/ #ComputerVision #AI #Robotics

Our NIST Task Board 1 has arrived courtesy of UMass Lowell's NERVE Center - thanks @UMLResearch! We're looking forward to many hours of happy assembly and disassembly 🦾🙂

Developing algorithms for autonomy in space is hard due to the lack of representative data. We present ALLO, a dataset of realistic synthetic images for proximity operations in lunar orbit. Open-source dataset and pipeline: github.com/utiasSTARS/ALL… Paper: arxiv.org/abs/2409.20435

We present placing via picking (PvP), a method to autonomously collect real-world demonstrations for placing tasks in which objects must be manipulated to contact-constrained locations. PhD candidate @OliverLimoyo will present it at #IROS2024! 🦾 Paper: arxiv.org/abs/2312.02352

Say cheese! We introduce PhotoBot, a framework for fully automated photo acquisition based on an interplay between high-level human language guidance and a robot photographer. PhD candidate @OliverLimoyo will present this work at #IROS2024! Paper: arxiv.org/abs/2401.11061