多重注意力和级联上下文糖网病病灶分割  

MCFNet:multi-attention and cascaded context fusion network for segmentation multiple lesion of diabetic retinopathy images

作　　者：郭燕飞 杜杭丽 杨成龙 孔祥真[1] Guo Yanfei;Du Hangli;Yang Chenglong;Kong Xiangzhen(College of Computer Science and Engineering,Qufu Normal University,Rizhao 276827,China)

机构地区：[1]曲阜师范大学计算机学院,日照276827

出　　处：《中国图象图形学报》2024年第12期3800-3816,共17页Journal of Image and Graphics

基　　金：国家自然科学基金项目(61976126);中国青年创新研究团队计划项目(2022KJ179);山东省自然科学基金项目(ZR2019MF003);日照市自然科学基金项目(RZ2022ZR64)。

摘　　要：目的 糖尿病视网膜病变(糖网病)(diabetic retinopathy,DR)是人类致盲的首要杀手,自动准确的糖网病病灶分割对于糖网病分级和诊疗至关重要。然而,不同类型的糖网病病灶结构复杂,大小尺度不一致且存在类间相似性和类内差异性,导致同时准确分割多种病灶充满挑战。针对上述问题,提出一种基于多重注意力和级联上下文融合的糖网病多类型病灶分割方法。方法 首先,三重注意力模块提取病灶的通道注意力、空间注意力和像素点注意力特征并进行加法融合以保证病灶特征的一致性。另外,级联上下文特征融合模块采用自适应平均池化和非局部操作提取不同层网络的全局上下文信息以扩大病灶的感受野。最后,平衡注意力模块计算病灶前景、背景和边界注意力图,并利用挤压激励模块在特征通道之间加权以重新平衡3个区域的注意力,令网络更多关注病灶的边缘细节,实现精细化分割。结果 在国际公开的糖网病图像数据集DDR(dataset for diabetic retinopathy)、IDRiD(Indian diabetic retinopathy image dataset)和E-Ophtha进行充分的对比实验和消融实验,4种病灶分割的平均AUC(area under curve)分别达到0.679 0、0.750 3和0.660 1。结论 基于多重注意力和级联上下文融合的糖网病分割方法(multi-attention and cascaded context fusion network,MCFNet)能够克服其他眼底组织和病灶噪声的不良干扰,同时实现糖网病4种病灶的精准分割,具有较好的准确性和鲁棒性,为临床医生进行糖网病诊疗提供有力支持。Objective Diabetic retinopathy(DR) is a leading cause of blindness in humans,and regular screening is helpful for its early detection and containment.While automated and accurate lesion segmentation is crucial for DR grading and diagnosis,this approach encounters many challenges due to the complex structures,inconsistent scales,and blurry edges of different kinds of lesions.However,the manual segmentation of DR lesions is time-consuming and labor-intensive,thus making the large-scale popularization of the approach particularly difficult due to the limited doctor resources and the high cost of manual annotation.Therefore,an automatic DR lesion segmentation method should be developed to reduce clinical workload and increase efficiency.Recently,convolutional neural networks have been widely applied in the fields of medical image segmentation and disease classification.The existing deep-learning-based methods for DR lesion segmentation are classified into image-based and patch-based approaches.Some studies have adopted the attention mechanism to segment lesions using the whole fundus image as input.However,these methods may lose the edge details of lesions,thus introducing challenges in obtaining fine-grained lesion segmentation results.Other studies have cropped the original images to patches and inputted them into the encoder-decoder networks for DR lesion segmentation.However,most of the approaches proposed in the literature utilize fixed weights to fuse coding features at different levels while ignoring the information differences among them,thus hindering the effective integration of multi-level features for accurate lesion segmentation.To address these issues,this paper proposes a multi-attention and cascaded context fusion network(MCFNet) for the simultaneous segmentation of multiple lesions.Method The proposed network adopts an encoder-decoder framework,including the VGG16 backbone network,triple attention module(TAM),cascaded context fusion module(CFM),and balanced attention module(BAM).First,directly fusing

关 键 词：糖尿病视网膜病变(DR) 多病灶分割 三重注意力 级联上下文融合 平衡注意力 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]