期刊
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
卷 84, 期 -, 页码 -出版社
ELSEVIER SCI LTD
DOI: 10.1016/j.bspc.2023.104810
关键词
Age -related macular degeneration; Retinal OCT image; Convolutional neural network; Class activation map; Image classification; Image segmentation
Age-related macular degeneration (AMD) is a common cause of visual impairment in the elderly, characterized by drusen and choroidal neovascularization (CNV). This paper proposes a deep learning framework that utilizes dual guidance for image classification and segmentation tasks in AMD diagnosis. The framework achieves higher accuracy in classification and segmentation compared to other networks tested on public datasets. The results also demonstrate the generalizability of the proposed model for detecting macular edema and segmentation of retinal fluids.
Age-related macular degeneration (AMD) is one of the main causes of visual impairment in elderly people, with drusen and choroidal neovascularization (CNV) being two characterizing types of lesions. Based on optical coherence tomography (OCT), image classification can be used in AMD diagnosis, while image segmentation is necessary for quantitative assessment of the lesion area. In this paper, we propose a deep learning framework exploiting dual guidance between the two tasks. Firstly, a complementary mask guided convolutional neural network (CM-CNN) is proposed to perform classification of OCT B-scans with drusen or CNV from normal ones, where the guiding mask is generated by the auxiliary segmentation task. Secondly, a class activation map guided UNet (CAM-UNet) is proposed to achieve segmentation of drusen and CNV lesions, using CAM output from the CM-CNN. Tested on a subset of the public UCSD dataset, and compared with five classification networks, four segmentation networks, and three multi-task networks, the proposed dual guidance network has achieved higher accuracy both in classification and segmentation. The classification accuracy reaches 96.93% and the Dice coefficient for segmentation reaches 77.51%. Results on an extra dataset for detection of macular edema and segmentation of retinal fluids further show the generalizability of the proposed model.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据