Skripsi

PENERAPAN MULTI PLANAR RECONSTRUCTION SLICING PADA ARSITEKTUR VISUAL GEOMETRY GROUP-16, UNET, DAN INCEPTION DALAM SEGMENTASI CITRA TIGA DIMENSI PADA HATI

---

Penilaian

---

The UNet architecture has been widely developed for image segmentation. UNet segmentation has been widely applied, one of which is on heart images. Segmentation of liver images is usually carried out on Abdominal Computed Tomography (CT) Scan images. The image from an abdominal CT scan is usually three-dimensional (3D) which focuses on organs including the liver, stomach, small intestine, large intestine, gallbladder, spleen, pancreas and kidneys. Unfortunately, 3D images are very complex and the U-Net architecture requires large amounts of data which can cause overvitting and suboptimal performance. A technique that can overcome the complexity of 3D images can be done using the MPR cutting technique. MPR slicing cuts 3D images into 2D based on the axial, sagittal and coronal axes. A new architectural combination is needed that can improve UNet performance. This research proposes a modification of the VGG-16, Inception, and UNet architecture. The VGG-16, Inception, and UNet architectures modify UNet by replacing the UNet encoder with a VGG-16 encoder to help extract more features. The VGG-16, Inception, and UNet architectures implement Inception on the bridge to help reduce the number of parameters. The application of the VGG-16, Inception, and UNet architecture to liver segmentation resulted in accuracy, sensitivity, specificity, F1-Score, and IoU values achieved above 95%. This shows that the performance of the VGG-16, Inception, and UNet architecture is very good in segmenting liver images. Based on the results of this performance evaluation, it can be concluded that the VGG-16, Inception, and UNet architectures are superior in segmenting liver CT scan images.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2407003774	2407003774	T1462772024	Central Library (references)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T1462772024

Publisher

Indralaya : Jurusan Matematika, Fakultas Matematika Dan Ilmu Pengetahuan Alam, Universitas Sriwijaya., 2024

Collation

ix, 52 hlm.; ilus.; tab.; 28 cm

Language

Indonesia

ISBN/ISSN

-

Classification

516.307

Content Type

Text

Media Type

unmediated

Carrier Type

unspecified

Edition

-

Subject(s)

Jurusan Matematika
VGG-16

Specific Detail Info

-

Statement of Responsibility

UIN Dimas

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment