Automatic Calcified Plaques Detection in the OCT Pullbacks Using Convolutional Neural Networks

作     者：Chunliu He Yifan Yin Jiaqiu Wang Biao Xu Zhiyong Li 

作者机构：The School of Biological Science and Medical EngineeringSoutheast UniversityNanjing210096China The Schoolof ChemistryPhysics and Mechanical EngineeringQueensland University of TechnologyBrisbaneQLD4000Australia Department of CardiologyAffiliated Drum Tower HospitalSchool of MedicineNanjing UniversityNanjing21000China 

出 版 物：《医用生物力学》 (Journal of Medical Biomechanics)

年 卷 期：2019年第34卷第A1期

页      面：109-110页

核心收录：

学科分类：0831[工学-生物医学工程（可授工学、理学、医学学位）] 100207[医学-影像医学与核医学] 1006[医学-中西医结合] 1002[医学-临床医学] 1001[医学-基础医学(可授医学、理学学位)] 08[工学] 1010[医学-医学技术（可授医学、理学学位）] 100201[医学-内科学(含：心血管病、血液病、呼吸系病、消化系病、内分泌与代谢病、肾病、风湿病、传染病)] 100106[医学-放射医学] 10[医学] 100602[医学-中西医结合临床] 

基　　金：supported in part by the National Natural Science Foundation of China ( NSFC ) ( 11772093) ARC ( FT140101152) 

主　　题：Calcified plaque intravascular optical coherence tomography deep learning imbalance label distribution convolutional neural networks 

摘      要：Background Coronary artery calcification is a well-known marker of atherosclerotic plaque ***-resolution intravascular optical coherence tomography(OCT)imaging has shown the potential to characterize the details of coronary calcification in *** routine clinical practice,it is a time-consuming and laborious task for clinicians to review the over 250 images in a single ***,the imbalance label distribution within the entire pullbacks is another problem,which could lead to the failure of the classifier *** the success of deep learning methods with other imaging modalities,a thorough understanding of calcified plaque detection using Convolutional Neural Networks(CNNs)within pullbacks for future clinical decision was *** All 33 IVOCT clinical pullbacks of 33 patients were taken from Affiliated Drum Tower Hospital,Nanjing University between December 2017 and December *** ground-truth annotation,three trained experts determined the type of plaque that was present in a *** experts assigned the labels no calcified plaque , calcified plaque for each OCT *** experts were provided the all images for *** final label was determined based on consensus between the experts,different opinions on the plaque type were resolved by asking the experts for a repetition of their *** the implement of algorithm,all OCT images was resized to a resolution of 300×300,which matched the range used with standard architectures in the natural image *** the study,we randomly selected 26 pullbacks for training,the remaining data were ***,imbalance label distribution within entire pullbacks was great challenge for various CNNs *** order to resolve the problem,we designed the following ***,we fine-tuned twenty different CNNs architecture,including customize CNN architectures and pretrained CNN *** the nature of OCT images,customize CNN architectures were des