Skripsi

PENGENALAN POLA PANDANGAN JANTUNG JANIN MENGGUNAKAN ARSITEKTUR FASTER R-CNN

---

Penilaian

---

Clinicians need the capabilities of deep learning to minimize healthcare costs, diagnose patient diseases, increase the chance to recover, and save many lifes. Deep learning has also proven to optimize the accuracy of diagnoses. The literatures of fetal heart are still limited due to heart screening views data can't be easily access by everyone. This study has explored four fetal heart screening views by ultrasongraphy videos. It consists of four chamber view, left ventricular outflow tract, right ventricular outflow tract, and three vessel and thrachea view. For pre-processing data, the videos have converted to frames. Also, the frames have annotated and augmented in training data. Faster R-CNN have implemented to get more accuracy and better optimization. This study used VGG16 and ResNet50 as backbone or feature extractors and pre-trained model from PASCAL VOC 2007. As the results, VGG16 outperformed ResNet50 model. Faster R-CNN with VGG16 achieves AP (average precison) for four chamber view 86.09%, left ventricular outflow tract 86.87%, right ventricular outflow tract 96.35%, and three vessel and trachea view 89.46%. For unseen data, the model gets mAP (mean average precision) 89.69%.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2107002544	T51046	T510462021	Central Library (Referens)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T510462021

Publisher

Inderalaya : Prodi Sistem Komputer, Fakultas Ilmu Komputer Uniersitas Sriwijaya., 2021

Collation

xxiii, 130 hlm,: ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

005.707

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Sistem Komputer
Data Sistem Komputer

Specific Detail Info

-

Statement of Responsibility

MURZ

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment