Skripsi

SEGMENTASI PEMBULUH DARAH PADA CITRA RETINA MENGGUNAKAN KOMBINASI ARSITEKTUR VISUAL GEOMETRY GROUP (VGG)-UNET

---

Penilaian

---

One of the processes in retinal image processing is retinal blood vessel segmentation. The retinal blood vessel segmentation process can use available datasets, one of which is the DRIVE dataset. The number of datasets and the size of the image used in the segmentation process will affect the performance of a method. One method that has strong ability when trained with large datasets is the Convolutional Neural Network (CNN). The main capability of CNN lies in its architecture where the oldest architecture that is often used is UNet, but UNet has a weakness that is it has a large number of parameters. This will have an impact on the execution time carried out to be be long. Another CNN-based architecture is the Visual Geometry Group (VGG). The VGG architecture has fewer parameters than UNet so that the execution time will be faster than UNet, but VGG is more often used in the image classification process than segmentation. The segmentation process carried out in this study using the combination of the advantages VGG and UNet architectures obtained quite good results. The results of the model performance obtained are 95% accuracy, 77% sensitivity, 96% specificity, 73% F-1 Score and 58% IoU. Based on these results, it shows that the combination of the VGG-UNet architecture in predicting retinal blood vessels and the results of black objects (background) is quite good, but fine retinal blood vessels have not been detected correctly.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2107003398	T56096	T560962021	Central Library (REFERENCES)	Available

Detail Information

Series Title

-

Call Number

T560962021

Publisher

Inderalaya : Prodi Ilmu Matematika, Fakultas Matematika dan Ilmu Pengetahuan Alam., 2021

Collation

xiii, 63 hlm. : ilus. ; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

516.07

Content Type

Text

Media Type

unmediated

Carrier Type

-

Edition

-

Subject(s)

geometry
Prodi Ilmu Matematika
Pembuluh Darah

Specific Detail Info

-

Statement of Responsibility

DS

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment