Smart Cameras for Smarter Autonomous Vehicles and Robots

This is the archived 1st edition. Visit the current 2nd Workshop on Smart Cameras for Physical AI (BMVC 2026).

Introduction

Welcome to the workshop on Smart Cameras for Smarter Autonomous Vehicles & Robots! This one-day event at BMVC 2025 brought together vision, robotics, and computational imaging experts to discuss challenges, applications, and innovations in embedded camera systems for robotic platforms.

Many advances in the development of autonomous robots and self-driving cars are limited by their embedded sensors. Arguably, the most important sensors in autonomous vehicles and robots are the cameras. The entire navigation and driving pipeline relies on high-quality visual input data. Vision-based algorithms for object recognition and scene understanding are key to ensuring proper navigation and interaction in the real world.

However, what happens if the camera gets damaged? What if the visual information is noisy? The safety and proper behavior of many autonomous systems depend on the quality and reliability of the cameras. For this reason, we introduced the 1st “Smart Cameras for Smarter Autonomous Vehicles and Robots” Workshop to unify low-level vision, computational photography, and robotics.

Schedule

Call for Papers [CLOSED]

We invited paper submissions on topics related to computational photography and low-level computer vision with applications in robotics and autonomous driving. Topics included, but were not limited to:

• camera‑based perception
• lens and optics
• image signal processors (ISP)
• image restoration
• drone and aerial photography
• novel image datasets
• sensor fusion
• real‑time processing
• vision language models (VLMs and VLAs)
• edge AI

Submission platform: https://cmt3.research.microsoft.com/SmartCameras2025/
Papers followed the BMVC 2025 format. Reviewing was double blind. Author Guidelines of BMVC 2025.

Important Dates