Scalable Pitch-Constrained Neural Processing Unit for 3D Integration with Event-Based Imagers

Abstract

Event-based imagers are bio-inspired sensors presenting intrinsic High Dynamic Range and High Acquisition Speed properties. However, noisy pixels and asynchronous readout result in poor energyefficiency and excessively large output data rates.In this work, we use Convolutional Spiking Neural Network filters to compensate these drawbacks and reduce output bandwidth by 10x.We designed a neuromorphic core as a distributable block that benefits from 3D integration technology with direct and parallel access to 32x32 pixels, enabling reduced frequency operation. Post-layout simulations depict a peak energy efficiency with 2.83pJ per Synaptic Operation (equivalent to 0.093fJ/event/pix) at the nominal literature input event rate.

Technological Evolution of Image Sensing Designed by Nanostructured Materials

“…Researchers are currently working to improve these qualities and illuminace to the level of 5 lx, which is a very gloomy environment. The human eye’s intrascene and interscene DR are normally 120 and 180 dB, respectively, following the contemporary smartphone’s intrascene and interscene DR of 70 and 120 dB, respectively, which is being improved constantly, whereas pixel size and resolution are the main determinants of these features . It is desirable that the pixel size be smaller to get a better CIS resolution.…”