针对肿瘤生长预测的边缘注入自适应时空网络  

Adaptive spatiotemporal networks based on edge injection for tumor growth prediction

作　　者：张晶[1,2,3] 李思阳 王河喜 ZHANG Jing;LI Siyang;WANG Hexi(School of Information Engineering and Automation,Kunming University of Technology,Kunming 650500,P.R.China;Yunnan Provincial Key Laboratory of Artificial Intelligence,Kunming 650500,P.R.China;Yunnan Provincial Key Laboratory of Computer Technology Application,Kunming 650500,P.R.China;School of Computer Science and Technology,Taiyuan University of Technology,Taiyuan 030000,P.R.China)

机构地区：[1]昆明理工大学信息工程与自动化学院,昆明650500 [2]云南省人工智能重点实验室,昆明650500 [3]云南省计算机技术应用重点实验室,昆明650500 [4]太原理工大学计算机科学与技术学院,太原030000

出　　处：《重庆邮电大学学报（自然科学版）》2024年第6期1194-1204,共11页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：云南省基础研究计划重点项目(2021AS070016);云南省重大科技专项计划项目(202302AD080002);云南省计算机技术应用重点实验室开放基金项目(FCB22144S073A);“兴滇英才支持计划”产业创新人才项目([2019]1096)。

摘　　要：为了更有效地表征肿瘤的生长趋势,提出了基于边缘注入的自适应时空网络(adaptive spatiotemporal networks based on edge injection,EA-Net)。引入紧密型空间增强模块学习多形态肿瘤的静态抽象信息,并重点提取关键生长特征;明确监督目标肿瘤的边缘信息,设计多密度边缘注入模块来增强特征图像内目标的边缘权重;建立广泛的时空信息关联,实现对肿瘤未来成像的自适应建模;引入个体特异性因子,促使网络学习不同人群肿瘤的生长趋势。实验表明,提出的方法能更高效地预测肿瘤未来的生长状态,其Dice score、Recall、RMSE分别可达89.36%、90.83%、10.42%,能在一定程度上揭示肺癌疾病的发展方向。information of multi-morphological tumors,focusing on extracting critical growth features.By explicitly supervising the target tumor’s edge information,a multi-density edge injection module is designed to enhance edge weights of the target within feature images.Broad spatiotemporal information associations are established to enable adaptive modeling of future tumor imaging.Individual-specific factors are incorporated to facilitate the network’s learning of tumor growth trends across diverse populations.Experimental results show that the proposed method can efficiently predict the future growth state of tumors,achieving a Dice score of 89.36%,Recall of 90.83%,and RMSE of 10.42%.This method provides insights into the progression of lung cancer to a certain extent.

关 键 词：肿瘤生长率 边缘注入 生长趋势 自适应预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TN919[自动化与计算机技术—计算机科学与技术]