Abstract:Optical convolution computing is gaining traction owing to its inherent parallelism, multi-dimensional processing, and energy efficiency. To handle input dimensions of N, conventional implementations necessitate N2 optical elements, such as Mach-Zehnder interferometers or micro-ring resonators, to process multiply-accumulate (MAC) operations, limiting scalability and resulting in elevated power consumption. Here, a direct convolution computing method based on wavelength routing, utilizing the unique sliding property of an arrayed waveguide grating router (AWGR) to perform the sliding window operation of the convolution in the wavelength-space domains is proposed. With two input vectors directly loaded onto two modulator arrays, the convolution result is instantaneously produced at a photodetector array. The entire convolution computation is executed within a single clock cycle without the need for preprocessing or decomposition into elementary MAC operations. The number of active elements is minimal, only needed for input/output. The proposed optical convolution unit has striking advantages of high scalability, high speed, and processing simplicity compared to those based on optical matrix-vector multipliers. In the first experimental demonstration, a remarkable classification accuracy of up to 98.2% in handwritten digit recognition tasks using a LeNet-5 neural network is achieved. A direct convolution computing method based on wavelength routing, utilizing the unique sliding property of an arrayed waveguide grating router (AWGR) to perform the sliding window operation of the convolution in the wavelength-space domains is proposed and demonstrated. With two inputs vectors directly loaded onto modulator arrays, the convolution result is instantaneously produced at a photodetector array. image

Gray-Tone Image-Processing by Threshold Superposition in an Optical Cellular-Logic Binary-Image Processor

Novel gray image processing in a single operation optical morphological processor

The Realization Of Parallel Subtraction Using Lclv In Optical Neural Network

Binary-Conversion of Image with Feature Preserving

Gray Scale Generation and Stabilization in Photo-Aligned Ferroelectric Liquid Crystal

Developed, Binary, Image Processing in a Dual-Channel, Optical, Real-Time Morphological Processor

The New Optical Scheme for Parallel-Processing of 1-D Gray Images

Direct Optical Convolution Computing Based on Arrayed Waveguide Grating Router

Optical scheme for parallelly realizing the morphological erosion of 1-D gray images

Programmable Real-Time Optoelectronic Hybrid Threshold Logic Processors

POLARIZATION-OPTICAL STACKED INTEGRATION SCHEME OF CELLULAR LOGIC 2-LAYER IMAGE-PROCESSOR

Hybrid optical-electronic implementation of morphological detection

Optical Implementation of Binary Image Morphological Transformations Using A Liquid-Crystal Light Valve (Lclv)

Reconfigurable Morphological Processor for Grayscale Image Processing

Optical Architecture for Implementing Image-Template Operations of Image Algebra

Progressive Color Visual Cryptography

Binary genetic algorithm-based pattern LUT for grayscale digital half-toning

Programmable Optical Binary Threshold Logic Implementation Via Optoelectronic Bistable Circuit

Parallel optical logic processor and bit slice-full adder using a single electron trapping device

A Biological Retina Inspired Tone Mapping Processor for High-Speed and Energy-Efficient Image Enhancement

Bitwise Operations of Cellular Automaton on Gray-scale Images