Image of SEGMENTASI LUBANG DAN RUANG JANTUNG ANAK NON DOPPLER MELALUI PERANGKAT MOBILE MENGGUNAKAN YOU ONLY LOOK ONCE (YOLO)

Skripsi

SEGMENTASI LUBANG DAN RUANG JANTUNG ANAK NON DOPPLER MELALUI PERANGKAT MOBILE MENGGUNAKAN YOU ONLY LOOK ONCE (YOLO)

Penilaian

0,0

dari 5
Penilaian anda saat ini :  

Non-Doppler segmentation of pediatric heart holes and chambers using deep learning approach with You Only Look Once (YOLO) algorithm through mobile devices. The main objective of this study is to evaluate the performance of YOLOv8 in segmenting cardiac structures, namely chambers and orifices, such as the right atrium, left atrium, right ventricle, left ventricle, and heart pit, from non-Doppler ultrasound images based on five main views: Four-chamber (4CH), Five-chamber (5CH), Long Axis (LA), Short Axis (SA), and Subcostal (SUB). Evaluation using mAP50 and IoU metrics on validation and unseen data showed that the best model, YOLOv8l with 200 epochs and batch size 4, achieved the highest segmentation performance on the 4CH view (mAP50: 0. 851, IoU: 0.709), 5CH (mAP50: 0.876, IoU: 0.763), LA (mAP50: 0.905, IoU: 0.754), SA (mAP50: 0.850, IoU: 0.724), and SUB (mAP50: 0.720, IoU: 0.697). Comparison with the raw data shows higher IoU performance, which is 0.787 in 4CH view, 0.798 in 5CH view, 0.802 in LA view, 0.829 in SA view as the best value, and 0.806 in SUB view. These results show that the implementation of YOLOv8 via mobile devices is quite good, even though the raw data shows superior segmentation performance.

Availability
Inventory Code Barcode Call Number Location Status
2507003942T177898T1778982025Central Library (Reference)Available but not for loan - Not for Loan
Detail Information
Series Title
-
Call Number
T1778982025
Publisher
Indralaya : Prodi Sistem Komputer, Fakultas Ilmu Komputer Universitas Sriwijaya.,
Collation
xvii, 117 hlm.; ilus.; tab.; 29 cm.
Language
Indonesia
ISBN/ISSN
-
Classification
004.07
Content Type
Text
Media Type
-
Carrier Type
-
Edition
-
Subject(s)
Sistem komputer
Specific Detail Info
-
Statement of Responsibility
Other version/related

No other version available

File Attachment
Comments
You must be logged in to post a comment