Fig. 4: In-sensor image transformations using the CIPS photodetector.
From: Ionic-electronic photodetector for vision assistance with in-sensor image processing
a Example of a highly noisy image where the signal object is completely obscured by noise, with a signal-to-noise (SNR) ratio of ~0.7. b Conventional image processing workflow: capturing the image with a camera, storing it in a memory device, and processing it using transformation functions in a computation unit. c In-sensor image processing workflow using the CIPS photodetector: performing image transformations directly during detection. Processed image using the computational algorithm (d) and CIPS photodetector (e). f Image transformations using the CIPS photodetector: the raw image, captured by a charge-coupled device camera in a dark scene, with signal area outlined by a white-dashed box, and the background area by a green-dashed box. The signal-to-background intensity ratio (SBR) increases from 0.14 to 1.24 as accumulation time increases from 0.5 to 10 s, dramatically enhancing visibility of objects in the dark. g Histogram of raw and processed images at different illumination times (t = 0.5 s, 5 s, and 10 s).
