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

Zhang，Xiaoqing1; Wu，Xiao1; Xiao，Zunjie1; Hu，Lingxi1; Qiu，Zhongxi1; Sun，Qingyang1; Higashita，Risa1,2; Liu，Jiang1,3,4

2023

10.1049/cit2.12246

CAAI Transactions on Intelligence Technology   影响因子和分区

2468-6557

2468-2322

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.

artificial intelligence deep learning deep neural networks image analysis image classification medical applications medical image processing

SCI ; EI

英语

第一 ; 通讯

Stable Support Plan Program[20200925174052004] ; Shenzhen Natural Science Fund[JCYJ20200109140820699] ; National Natural Science Foundation of China[82272086] ; Guangdong Provincial Department of Education["2020ZDZX3043","SJZLGC202202"] ; Guangdong Provincial Key Laboratory[2020B121201001]

Computer Science

Computer Science, Artificial Intelligence

WOS:001000135500001

WILEY

20232414216149

Complex networks ; Convolution ; Convolutional neural networks ; Diagnosis ; Economic and social effects ; Image classification ; Medical applications ; Medical imaging ; Network architecture ; Optical data processing ; Optical tomography ; Parameter estimation ; Topology

Biomedical Engineering:461.1 ; Ergonomics and Human Factors Engineering:461.4 ; Medicine and Pharmacology:461.6 ; Information Theory and Signal Processing:716.1 ; Computer Systems and Equipment:722 ; Data Processing and Image Processing:723.2 ; Optical Devices and Systems:741.3 ; Imaging Techniques:746 ; Combinatorial Mathematics, Includes Graph Theory, Set Theory:921.4 ; Social Sciences:971

2-s2.0-85161382657

Scopus

1.Research Institute of Trustworthy Autonomous Systems and Department of Computer Science and Engineering,Southern University of Science and Technology,Shenzhen,China
2.Tomey Corporation,Nagoya,Japan
3.Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation,Department of Computer Science and Engineering,Southern University of Science and Technology,Shenzhen,China
4.Singapore Eye Research Institute,Singapore,Singapore

Zhang，Xiaoqing,Wu，Xiao,Xiao，Zunjie,et al. Mixed-decomposed convolutional network: A lightweight yet efficient convolutional neural network for ocular disease recognition[J]. CAAI Transactions on Intelligence Technology,2023.

Zhang，Xiaoqing.,Wu，Xiao.,Xiao，Zunjie.,Hu，Lingxi.,Qiu，Zhongxi.,...&Liu，Jiang.(2023).Mixed-decomposed convolutional network: A lightweight yet efficient convolutional neural network for ocular disease recognition.CAAI Transactions on Intelligence Technology.

Zhang，Xiaoqing,et al."Mixed-decomposed convolutional network: A lightweight yet efficient convolutional neural network for ocular disease recognition".CAAI Transactions on Intelligence Technology (2023).