Bio-Inspired Camera Mimicking the Structural and Functional Features of Biological Vision

Abstract

Conventional imaging and data processing devices may not be ideal for mobile machine vision applications, such as drones and robots, due to the bulky and heavy multi-lens optics used in conventional cameras. Additionally, physical isolation of camera and processors necessitates the capture, transfer, and processing of redundant data, resulting in large power consumption and data latency. Here, we review on bio-inspired cameras inspired by the structural and functional features of biological eyes. We first summarize recent strategies to fabricate curved image sensor (CurvIS) arrays mimicking the curved retina of biological eyes. These CurvIS arrays enable aberration-free imaging with a single lens optics, leading to a miniaturization of camera module. Other optical advantages, such as wide field-of-view and deep depth-of-field, could be also offered by the integrated camera module. Next, we discuss bio-inspired cameras capable of performing in-sensor processing as well as image acquisition. As notable examples, we introduce synaptic optoelectronic devices that can efficiently enhance image contrast and reduce the noise using photon-triggered synaptic plasticity.