Exposure-Programmable CMOS Pixel With Selective Charge Storage and Code Memory for Computational Imaging

Computational imaging, as a rapid emerging field, enables unprecedented photographic capability through the encoding of visual information. Currently, as exposure programming is not support by image sensors, computational cameras use off-chip components to accomplish exposure/encoding optical computations. In this paper, we present a complementary metal-oxide-semiconductor (CMOS) pixel structure with exposure encoding capabilities that enable on-chip computational imaging. As a proof-of-concept, a 10 x 10-pixel prototype chip is fabricated in a 0.13-mu m 8-metal 1-poly CMOS process with shared pixel output wires and two 3.3-V power supplies. Each computational pixel is 12.1 mu m x 12.2 mu m with a fill factor of 33.2%. Measurement results confirm that at a frame rate of 60 f/s, the proposed pixel is capable of performing both temporal and spatial exposure encoding. The proposed pixel design paves a promising path to achieve on-chip temporal-spatial exposure encoding directly on the sensor focal plane for computational imaging.