Mixed-decomposed convolutional network: A lightweight yet efficient convolutional neural network for ocular disease recognition

被引:10
|
作者
Zhang, Xiaoqing [1 ,2 ,6 ,7 ]
Wu, Xiao [1 ,2 ]
Xiao, Zunjie [1 ,2 ]
Hu, Lingxi [1 ,2 ]
Qiu, Zhongxi [1 ,2 ]
Sun, Qingyang [1 ,2 ]
Higashita, Risa [1 ,2 ,3 ,6 ,7 ]
Liu, Jiang [1 ,2 ,4 ,5 ,6 ,7 ]
机构
[1] Southern Univ Sci & Technol, Res Inst Trustworthy Autonomous Syst, Shenzhen, Peoples R China
[2] Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen, Peoples R China
[3] Tomey Corp, Nagoya, Japan
[4] Southern Univ Sci & Technol, Dept Comp Sci & Engn, Guangdong Prov Key Lab Brain inspired Intelligent, Shenzhen, Peoples R China
[5] Singapore Eye Res Inst, Singapore, Singapore
[6] Southern Univ Sci & Technol, Res Inst Trustworthy Autonomous Syst, Shenzhen 518055, Peoples R China
[7] Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen 518055, Peoples R China
基金
中国国家自然科学基金;
关键词
artificial intelligence; deep learning; deep neural networks; image analysis; image classification; medical applications; medical image processing; CLASSIFICATION;
D O I
10.1049/cit2.12246
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Eye health has become a global health concern and attracted broad attention. Over the years, researchers have proposed many state-of-the-art convolutional neural networks (CNNs) to assist ophthalmologists in diagnosing ocular diseases efficiently and precisely. However, most existing methods were dedicated to constructing sophisticated CNNs, inevitably ignoring the trade-off between performance and model complexity. To alleviate this paradox, this paper proposes a lightweight yet efficient network architecture, mixed-decomposed convolutional network (MDNet), to recognise ocular diseases. In MDNet, we introduce a novel mixed-decomposed depthwise convolution method, which takes advantage of depthwise convolution and depthwise dilated convolution operations to capture low-resolution and high-resolution patterns by using fewer computations and fewer parameters. We conduct extensive experiments on the clinical anterior segment optical coherence tomography (AS-OCT), LAG, University of California San Diego, and CIFAR-100 datasets. The results show our MDNet achieves a better trade-off between the performance and model complexity than efficient CNNs including MobileNets and MixNets. Specifically, our MDNet outperforms MobileNets by 2.5% of accuracy by using 22% fewer parameters and 30% fewer computations on the AS-OCT dataset.
引用
收藏
页码:319 / 332
页数:14
相关论文
共 50 条
  • [21] Image recognition based on lightweight convolutional neural network: Recent advances
    Liu, Ying
    Xue, Jiahao
    Li, Daxiang
    Zhang, Weidong
    Chiew, Tuan Kiang
    Xu, Zhijie
    IMAGE AND VISION COMPUTING, 2024, 146
  • [22] A Lightweight Fully Convolutional Neural Network for SAR Automatic Target Recognition
    Yu, Jimin
    Zhou, Guangyu
    Zhou, Shangbo
    Yin, Jiajun
    REMOTE SENSING, 2021, 13 (15)
  • [23] Improved Lightweight Convolutional Neural Network for Finger Vein Recognition System
    Hsia, Chih-Hsien
    Ke, Liang-Ying
    Chen, Sheng-Tao
    BIOENGINEERING-BASEL, 2023, 10 (08):
  • [24] Traffic signs andmarkings recognition based on lightweight convolutional neural network
    Zhao, Shuen
    Gong, Zhikun
    Zhao, Dongyu
    VISUAL COMPUTER, 2024, 40 (02): : 559 - 570
  • [25] Aerial Infrared Target Recognition Based on Lightweight Convolutional Neural Network
    Cui, Tingting
    Tang, Linbo
    Heng, Yong
    Li, Zhenzhen
    Nan, Jinghong
    ELEVENTH INTERNATIONAL CONFERENCE ON DIGITAL IMAGE PROCESSING (ICDIP 2019), 2019, 11179
  • [26] Recognition of fractional vortex optical modes with a lightweight convolutional neural network
    Xu, Fei
    Sheng, Tianze
    Liu, Lifeng
    Li, Xiaolin
    Qi, Yihong
    Qian, Jun
    Niu, Yueping
    Gong, Shangqing
    LASER PHYSICS, 2025, 35 (01)
  • [27] DCFNet: Deep Neural Network with Decomposed Convolutional Filters
    Qiu, Qiang
    Cheng, Xiuyuan
    Calderbank, Robert
    Sapiro, Guillermo
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 80, 2018, 80
  • [28] Lightweight convolutional neural network for vehicle recognition in thermal infrared images
    Kang, Qing
    Zhao, Hongdong
    Yang, Dongxu
    Ahmed, Hafiz Shehzad
    Ma, Juncheng
    INFRARED PHYSICS & TECHNOLOGY, 2020, 104
  • [29] Underwater target recognition using a lightweight asymmetric convolutional neural network
    Yan, Chenhong
    Yao, Tianyi
    Yu, Yang
    Yang, Changsheng
    Liu, Lu
    Pan, Guang
    Yan, Shefeng
    17TH ACM INTERNATIONAL CONFERENCE ON UNDERWATER NETWORKS & SYSTEMS, WUWNET 2023, 2024,
  • [30] Traffic signs and markings recognition based on lightweight convolutional neural network
    Shuen Zhao
    Zhikun Gong
    Dongyu Zhao
    The Visual Computer, 2024, 40 (2) : 559 - 570