A Silicon Image Sensor That Computes

New approach for image analysis?

A Silicon Image Sensor That Computes

Harvard University John A. Paulson School of Engineering and Applied Sciences

Leah Burrows, August 25, 2022

U.S. and South Korean researchers have developed in-sensor processors that could be incorporated into commercial complementary metal-oxide-semiconductor (CMOS) image sensors. The researchers created an electrostatically doped silicon photodiode array, enabling voltage-based tuning of pixels' optical sensitivity. Harvard University's Houk Jang said, "These dynamic photodiodes can concurrently filter images as they are captured, allowing for the first stage of vision processing to be moved from the microprocessor to the sensor itself." Users can program the array into different image filters to eliminate unnecessary details or noise for various applications. Harvard's Henry Hinton predicted the processor will see use "not only in machine vision applications, but also in bio-inspired applications, wherein early information processing allows for the co-location of sensor and compute units, like in the brain."