Optimization of optical convolution kerne of optoelectronic hybrid convolution neural network

被引：1
|
作者：
Xu Xiaofeng [1 ]
Zhu Lianqing [1 ,2 ]
Zhuang Wei [2 ]
Zhang Dongliang [2 ]
Lu Lidan [3 ]
Yuan Pei [3 ]
机构：
[1] Changchun Univ Sci & Technol, Sch Electroopt Engn, Changchun 130022, Jilin, Peoples R China
[2] Beijing Informat Sci & Technol Univ, Key Lab, Minist Educ Optoelect Measurement Technol & Instr, Beijing 100192, Peoples R China
[3] Beijing Informat Sci & Technol Univ, Beijing Lab Opt Fiber Sensing & Syst, Beijing 100016, Peoples R China
来源：
OPTOELECTRONICS LETTERS | 2022年 / 18卷 / 03期
基金：
中国国家自然科学基金;
关键词：
A;
D O I：
10.1007/s11801-022-1183-x
中图分类号：
O43 [光学];
学科分类号：
070207 ; 0803 ;
摘要：
To enhance the optical computation's utilization efficiency, we develop an optimization method for optical convolution kernel in the optoelectronic hybrid convolution neural network (OHCNN). To comply with the actual calculation process, the convolution kernel is expanded from single-channel to two-channel, containing positive and negative weights. The Fashion-MNIST dataset is used to test the network architecture's accuracy, and the accuracy is improved by 7.5% with the optimized optical convolution kernel. The energy efficiency ratio (EER) of two-channel network is 46.7% higher than that of the single-channel network, and it is 2.53 times of that of traditional electronic products.
引用
收藏
页码：181 / 186
页数：6
相关论文
共 50 条
  • [1] Optimization of optical convolution kernel of optoelectronic hybrid convolution neural network
    XU Xiaofeng
    ZHU Lianqing
    ZHUANG Wei
    ZHANG Dongliang
    LU Lidan
    YUAN Pei
    [J]. Optoelectronics Letters, 2022, (03) : 181 - 186
  • [2] Optimization of optical convolution kernel of optoelectronic hybrid convolution neural network
    Xiaofeng Xu
    Lianqing Zhu
    Wei Zhuang
    Dongliang Zhang
    Lidan Lu
    Pei Yuan
    [J]. Optoelectronics Letters, 2022, 18 : 181 - 186
  • [3] Infrared object classification with a hybrid optical convolution neural network
    Chen, Jianbo
    Talley, Jennifer
    Kelly, Kevin F.
    [J]. APPLIED OPTICS, 2021, 60 (25) : G224 - G231
  • [4] Compact lensless convolution processor for an optoelectronic convolutional neural network
    Zhang, Zaikun
    Kong, Depeng
    Da, Zhengshang
    Wang, Ruiduo
    Wang, Shijie
    Geng, Yi
    He, Zhengquan
    [J]. JOURNAL OF PHYSICS D-APPLIED PHYSICS, 2023, 56 (35)
  • [5] Hybrid optical convolutional neural network with convolution kernels trained in the spatial domain
    Gene, Jinhwa
    Park, Suntak
    Shin, Hyung Cheol
    Sohn, Jong Moo
    [J]. NEUROCOMPUTING, 2024, 573
  • [6] Optimization of Convolution Neural Network Algorithm Based on FPGA
    Tang, Feixue
    Zhang, Weichao
    Tian, Xiaogang
    Fan, Xiaoye
    Cao, Xixin
    [J]. EMBEDDED SYSTEMS TECHNOLOGY, ESTC 2017, 2018, 857 : 131 - 140
  • [7] Memory optimization at Edge for Distributed Convolution Neural Network
    Naveen, Soumyalatha
    Kounte, Manjunath R.
    [J]. TRANSACTIONS ON EMERGING TELECOMMUNICATIONS TECHNOLOGIES, 2022, 33 (12):
  • [8] Single Image Dehazing Using Hybrid Convolution Neural Network
    Juneja, Akshay
    Kumar, Vijay
    Singla, Sunil Kumar
    [J]. MULTIMEDIA TOOLS AND APPLICATIONS, 2023, 83 (13) : 38355 - 38386
  • [9] Single Image Dehazing Using Hybrid Convolution Neural Network
    Akshay Juneja
    Vijay Kumar
    Sunil Kumar Singla
    [J]. Multimedia Tools and Applications, 2024, 83 : 38355 - 38386
  • [10] Hybrid Particle Swarm training for Convolution Neural Network (CNN)
    Chhabra, Yoshika
    Varshney, Sanchit
    Ankita
    [J]. 2017 TENTH INTERNATIONAL CONFERENCE ON CONTEMPORARY COMPUTING (IC3), 2017, : 381 - 383
12345