Visual information plays a central role in how both humans and machines perceive and understand their environment. As the field of computer vision advances, various systems, such as autonomous vehicles, AR headsets, and robots, use multiple cameras to interpret the world. However, conventional cameras, built with refractive optics and CMOS image sensors, are optimized for imaging quality rather than for computer vision performance.

Our research aims to design engineered optics and sensors that are specifically optimized for vision tasks, addressing critical challenges such as latency and energy consumption. By performing part of the computation directly in the optical domain, our approach significantly reduces the burden on digital processing. We are also exploring quantum optical methods for pre-processing information, opening new possibilities for efficient and intelligent optical computing.