In-sensor image memorization and encoding via optical neurons for bio-stimulus domain reduction toward visual cognitive processing

• DOI: 10.1038/s41467-022-32790-3
• Publication Date: 2022-09-05T14:03:04Z
• Authors: Doeon Lee; Minseong Park; Yongmin Baek; Byungjoon Bae; Junseok Heo; Kyusang Lee
• Keywords: Neuromorphic engineering; Memristor

As machine vision technology generates large amounts of data from sensors, it requires efficient computational systems for visual cognitive processing. Recently, in-sensor computing have emerged as a potential solution reducing unnecessary transfer and realizing fast energy-efficient However, they still lack the capability to process stored images directly within sensor. Here, we demonstrate heterogeneously integrated 1-photodiode 1 memristor (1P-1R) crossbar processing, emulating mammalian image encoding extract features input images. Unlike other neuromorphic processes, trained weight values are applied an voltage image-saved array instead storing value in memristors, paradigm. We believe platform provides advanced architecture real-time data-intensive machine-vision applications via bio-stimulus domain reduction.