Comparative Study of Deep Learning Modelsfor Automatic Coronary Stenosis Detectionin X-ray Angiography; CEUR Workshop Proceedings; Vol. 2744 : GraphiCon 2020. Computer Graphics and Machine Vision

Détails bibliographiques
Parent link:CEUR Workshop Proceedings: Online Proceedings for Scientific Conferences and Workshops
Vol. 2744 : GraphiCon 2020. Computer Graphics and Machine Vision.— 2020.— [11 p.]
Collectivité auteur: Национальный исследовательский Томский политехнический университет Инженерная школа информационных технологий и робототехники Отделение автоматизации и робототехники
Autres auteurs: Danilov V. V. Vyacheslav Vladimirovich, Gerget O. M. Olga Mikhailovna, Klyshnikov K. Yu. Kirill Yurjevich, Ovcharenko E. A. Evgeny Andreevich, Frangi A. Alejandro
Résumé:Title screen
The article explores the application of machine learning approach to detect both single-vessel and multivessel coronary artery disease from X-ray angiography. Since the interpretation of coronary angiography images requires interventional cardiologists to have considerable training, our study is aimed at analysing, training, and assessing the potential of the existing object detectors for classifying and detecting coronary artery stenosis using angiographic imaging series. 100 patients who underwent coronary angiography at the Research Institute for Complex Issues of Cardiovascular Diseases were retrospectively enrolled in the study. To automate the medical data analysis, we examined and compared three models (SSD MobileNet V1, Faster-RCNN ResNet-50 V1, FasterRCNN NASNet) with various architecture, network complexity, and a number of weights. To compare developed deep learning models, we used the mean Average Precision (mAP) metric, training time, and inference time. Testing results show that the training/inference time is directly proportional to the model complexity. Thus, Faster-RCNN NASNet demonstrates the slowest inference time. Its mean inference time per one image made up 880 ms. In terms of accuracy, FasterRCNN ResNet-50 V1 demonstrates the highest prediction accuracy. This model has reached the mAP metric of 0.92 on the validation dataset. SSD MobileNet V1 has demonstrated the best inference time with the inference rate of 23 frames per second.
Langue:anglais
Publié: 2020
Collection:Session 11: Artificial Intelligence, Cognitive Technology and Robotics
Sujets:
Accès en ligne:http://ceur-ws.org/Vol-2744/paper75.pdf
Format: Électronique Chapitre de livre
KOHA link:https://koha.lib.tpu.ru/cgi-bin/koha/opac-detail.pl?biblionumber=662983

MARC

LEADER 00000naa0a2200000 4500
001 662983
005 20251127122401.0
035 |a (RuTPU)RU\TPU\network\34151 
035 |a RU\TPU\network\32702 
090 |a 662983 
100 |a 20210118d2020 k||y0engy50 ba 
101 0 |a eng 
102 |a DE 
135 |a drcn ---uucaa 
181 0 |a i  
182 0 |a b 
200 1 |a Comparative Study of Deep Learning Modelsfor Automatic Coronary Stenosis Detectionin X-ray Angiography  |f V. V. Danilov, O. M. Gerget, K. Yu. Klyshnikov[et al.] 
203 |a Text  |c electronic 
225 1 |a Session 11: Artificial Intelligence, Cognitive Technology and Robotics 
300 |a Title screen 
320 |a [References: 19 tit.] 
330 |a The article explores the application of machine learning approach to detect both single-vessel and multivessel coronary artery disease from X-ray angiography. Since the interpretation of coronary angiography images requires interventional cardiologists to have considerable training, our study is aimed at analysing, training, and assessing the potential of the existing object detectors for classifying and detecting coronary artery stenosis using angiographic imaging series. 100 patients who underwent coronary angiography at the Research Institute for Complex Issues of Cardiovascular Diseases were retrospectively enrolled in the study. To automate the medical data analysis, we examined and compared three models (SSD MobileNet V1, Faster-RCNN ResNet-50 V1, FasterRCNN NASNet) with various architecture, network complexity, and a number of weights. To compare developed deep learning models, we used the mean Average Precision (mAP) metric, training time, and inference time. Testing results show that the training/inference time is directly proportional to the model complexity. Thus, Faster-RCNN NASNet demonstrates the slowest inference time. Its mean inference time per one image made up 880 ms. In terms of accuracy, FasterRCNN ResNet-50 V1 demonstrates the highest prediction accuracy. This model has reached the mAP metric of 0.92 on the validation dataset. SSD MobileNet V1 has demonstrated the best inference time with the inference rate of 23 frames per second. 
461 |t CEUR Workshop Proceedings  |o Online Proceedings for Scientific Conferences and Workshops 
463 |t Vol. 2744 : GraphiCon 2020. Computer Graphics and Machine Vision  |o Proceedings of the 30th International Conference, Saint Petersburg, Russia, September 22-25, 2020  |v [11 p.]  |d 2020 
610 1 |a электронный ресурс 
610 1 |a труды учёных ТПУ 
610 1 |a stenosis detection 
610 1 |a X-ray angiography 
610 1 |a deep learning 
610 1 |a transfer 
610 1 |a learning 
610 1 |a ангиография 
610 1 |a глубокое обучение 
701 1 |a Danilov  |b V. V.  |c specialist in the field of informatics and computer technology  |c engineer of Tomsk Polytechnic University  |f 1989-  |g Vyacheslav Vladimirovich  |3 (RuTPU)RU\TPU\pers\37831 
701 1 |a Gerget  |b O. M.  |c Specialist in the field of informatics and computer technology  |c Professor of Tomsk Polytechnic University, Doctor of Sciences  |f 1974-  |g Olga Mikhailovna  |3 (RuTPU)RU\TPU\pers\31430  |9 15593 
701 1 |a Klyshnikov  |b K. Yu.  |g Kirill Yurjevich 
701 1 |a Ovcharenko  |b E. A.  |g Evgeny Andreevich 
701 1 |a Frangi  |b A.  |g Alejandro 
712 0 2 |a Национальный исследовательский Томский политехнический университет  |b Инженерная школа информационных технологий и робототехники  |b Отделение автоматизации и робототехники  |3 (RuTPU)RU\TPU\col\23553 
801 2 |a RU  |b 63413507  |c 20210118  |g RCR 
856 4 |u http://ceur-ws.org/Vol-2744/paper75.pdf 
942 |c CF