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

作     者：Xiaoqing Zhang Xiao Wu Zunjie Xiao Lingxi Hu Zhongxi Qiu Qingyang Sun Risa Higashita Jiang Liu 

作者机构：Research Institute of Trustworthy Autonomous Systems and Department of Computer Science and EngineeringSouthern University of Science and TechnologyShenzhenChina Tomey CorporationNagoyaJapan Guangdong Provincial Key Laboratory of Brain‐inspired Intelligent ComputationDepartment of Computer Science and EngineeringSouthern University of Science and TechnologyShenzhenChina Singapore Eye Research InstituteSingaporeSingapore 

出 版 物：《CAAI Transactions on Intelligence Technology》 (智能技术学报（英文）)

年 卷 期：2024年第9卷第2期

页      面：319-332页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Stable Support Plan Program,Grant/Award Number:20200925174052004 Shenzhen Natural Science Fund,Grant/Award Number:JCYJ20200109140820699 National Natural Science Foundation of China,Grant/Award Number:82272086 Guangdong Provincial Department of Education,Grant/Award Numbers:2020ZDZX3043,SJZLGC202202 Guangdong Provincial Key Laboratory,Grant/Award Number:2020B121201001。 

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

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,mixeddecomposed 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.