Text

KOMBINASI ARSITEKTUR U-NET VGG-19 DENGAN ATTENTION GATE DALAM SEGMENTASI SEMANTIK PEMBULUH DARAH ARTERI DAN VENA PADA CITRA RETINA

---

Penilaian

---

Retinal blood vessels are categorized into arteries and veins. Arteries and veins have distinct functions and characteristics. Separation of arteries and veins in retinal images by semantic segmentation is an important step in supporting the diagnosis process of various diseases. Differences in the characteristics of arteries and veins can indicate the presence of disorders in the retina. This research uses a combination of U-Net architecture with the addition of VGG-19 and Attention Gate to segment arterial and venous blood vessels in retinal images. VGG-19 is applied in all parts of the convolution block contained in the encoder section aimed at learning more complex images. Attention Gate is inserted in the skip connections to focus the model on relevant features. The results of the application of the proposed architecture resulted in average performance on accuracy, sensitivity, specificity, f1-score, and IoU are good in segmenting arterial and venous blood vessels with 98.61%, 81.72%, 91.34%, 82.43% 72.67%. The average performance on the background label shows that the accuracy, sensitivity, f1-score, and IoU values have achieved good performance above 90%, although the specificity is still at 75%. Meanwhile, on vein labels, accuracy and specificity show good performance with values above 90%. However, the performance on sensitivity, f1-score, and IoU is already quite good at above 70%. However, the arterial label is still low due to the relatively small size of arterial features and is difficult to recognize. It is necessary to improve this architecture to get sensitivity, f1-score, and IoU values above 90%.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2507001054	T167335	T1673352025	Central Library (Reference)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T1673352025

Publisher

: Prodi Ilmu Matematika, Fakultas Matematika Dan Ilmu Pengetahuan Alam Universitas Sriwijaya., 2025

Collation

xii, 88 hlm.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

510.07

Content Type

-

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Matematika

Specific Detail Info

-

Statement of Responsibility

EM

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment