Key points are not available for this paper at this time.
A dynamic vision sensor (DVS) encodes temporal contrast (TC) of light intensity into address-events that are asynchronously transmitted for subsequent processing. This paper describes a DVS with improved TC sensitivity and event encoding. To enhance the TC sensitivity, each pixel employs a common-gate photoreceptor for low output noise and a capacitively-coupled programmable gain amplifier for continuous-time signal amplification without sacrificing the intra-scene dynamic range. A proposed in-pixel asynchronous delta modulator (ADM) better preserves signal integrity in event encoding compared with self-timed reset (STR) used in previous DVSs. A 60 × 30 prototype sensor array with a 31.2 μm pixel pitch was fabricated in a 1P6M 0.18 μm CMOS technology. It consumes 720 μW at a 100k event/s output rate. Measurements show that a 1% TC sensitivity with a 35% relative standard deviation is achieved and that the in-pixel ADM is up to 3.5 times less susceptible to signal loss than STR in terms of event number. These improvements can facilitate the application of DVSs in areas like optical neuroimaging which is demonstrated in a simulated experiment.
Yang et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: