Machine Vision With InP Based Floating-Gate Photo-Field-Effective Transistors for Color-Mixed Image Recognition
IEEE Journal of Quantum Electronics(2022)
摘要
The success of artificial neural networks (ANNs) in machine vision techniques has driven hardware researchers to explore more efficient computing elements for energy-expensive operations such as vector-matrix multiplication (VMM). In this work, InP-based floating-gate photo-field-effective transistors (FG-PFETs) are demonstrated as computing elements that integrate both photodetection and initial signal processing at the sensor level. These devices are fabricated from semiconductor channels grown via a back-end CMOS compatible templated-liquid phase (TLP) approach. Individual devices are shown to exhibit programmable responsivity, mimicking the effect of a synapse connecting the photodetector to a neuron. Using these devices, a simulated optical neural network (ONN) where the experimentally measured performance of FG-PFETs is used as an input shows excellent image recognition accuracy for color-mixed handwritten digits.
更多查看译文
关键词
Machine vision,optical neural network,phototransistors,indium phosphide
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要