Automatic Fetal Motion Detection from Trajectory of US Videos Based on YOLOv5 and LSTM
dc.authorid | 0000 0002 4370 7474 | |
dc.authorid | 0000-0001-6559-1399 | |
dc.contributor.author | Turkan, Musa | |
dc.contributor.author | Urfalı, Furkan Ertürk | |
dc.contributor.author | Dandıl, Emre | |
dc.date.accessioned | 2025-04-28T17:45:12Z | |
dc.date.available | 2025-04-28T17:45:12Z | |
dc.date.issued | 2023 | en_US |
dc.department | Enstitüler, Fen Bilimleri Enstitüsü, Elektronik ve Bilgisayar Mühendisliği | |
dc.department | Fakülteler, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü | |
dc.description.abstract | Deep learning methods have been widely used in the processing and evaluation of medical data in recent years. Computer-aided systems, which are used as a tool in the analysis of medical images and videos, can facilitate the diagnosis process and contribute to increasing the accuracy in the decision-making stages of the experts. In addition, the evaluation of medical data, which requires experience and expertise, is achieved with the help of deep learning, providing convenience in diagnosis and treatment. In this chapter, a hybrid deep learning method based on YOLOv5 and LSTM algorithms is proposed for recognizing the anatomical structures of the fetus and detecting its movements using fetal ultrasound (US) videos. In the study, a dataset is prepared from ultrasound videos containing fetal movements. At the next stage, the anatomical structures of the fetus are determined on the labeled data and the movements are tracked. In the experimental analyses, the movements of the anatomical structures such as the heart, head and body of the fetus are tracked and motion trajectory patterns are extracted. In the last stage, the detection and classification of fetal anatomical structures are achieved with the LSTM deep learning algorithm, using movement patterns converted to a two-dimensional plane. | en_US |
dc.identifier.citation | Turkan, M., Urfalı, F. E., & Dandıl, E. (2023). Automatic Fetal Motion Detection from Trajectory of US Videos Based on YOLOv5 and LSTM. In Explainable Machine Learning for Multimedia Based Healthcare Applications (pp. 1-20). Cham: Springer International Publishing. | en_US |
dc.identifier.doi | 10.1007/978-3-031-38036-5_1 | |
dc.identifier.endpage | 20 | en_US |
dc.identifier.scopus | 2-s2.0-85194334900 | |
dc.identifier.scopusquality | N/A | |
dc.identifier.startpage | 1 | en_US |
dc.identifier.uri | https://doi.org/10.1007/978-3-031-38036-5_1 | |
dc.identifier.uri | https://hdl.handle.net/11552/3930 | |
dc.indekslendigikaynak | Scopus | |
dc.institutionauthor | Turkan, Musa | |
dc.institutionauthor | Dandıl, Emre | |
dc.language.iso | en | |
dc.publisher | Springer | en_US |
dc.relation.ispartof | Explainable Machine Learning for Multimedia Based Healthcare Applications | |
dc.relation.publicationcategory | Kitap Bölümü - Uluslararası | en_US |
dc.rights | info:eu-repo/semantics/closedAccess | |
dc.title | Automatic Fetal Motion Detection from Trajectory of US Videos Based on YOLOv5 and LSTM | |
dc.type | Book Chapter |