PENERAPAN MULTIPLANAR RECONSTRUCTION PADA ARSITEKTUR U-RESNET DAN MOBILENET DALAM PROSES SEGMENTASI HATI CITRA TIGA DIMENSI HASIL CT SCAN

IFFAH'DA, ANNISA NURBA and Desiani, Anita and Eliyati, Ning (2024) PENERAPAN MULTIPLANAR RECONSTRUCTION PADA ARSITEKTUR U-RESNET DAN MOBILENET DALAM PROSES SEGMENTASI HATI CITRA TIGA DIMENSI HASIL CT SCAN. Undergraduate thesis, Sriwijaya University.

	Text RAMA_44201_08011282025039.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_44201_08011282025039_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (6MB) | Request a copy
	Text RAMA_44201_08011282025039_0011127702_0020115903_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (1MB)
	Text RAMA_44201_08011282025039_0011127702_0020115903_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (772kB) | Request a copy
	Text RAMA_44201_08011282025039_0011127702_0020115903_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (314kB) | Request a copy
	Text RAMA_44201_08011282025039_0011127702_0020115903_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (882kB) | Request a copy
	Text RAMA_44201_08011282025039_0011127702_0020115903_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (175kB) | Request a copy
	Text RAMA_44201_08011282025039_0011127702_0020115903_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (423kB) | Request a copy

Abstract

U-Net is commonly used for medical image segmentation, such as liver organ segmentation. One tool that can visualize the morphology, shape, and position of the liver is a 3D Abdominal Computed Tomography (CT) Scan. However, U-Net requires a large amount of data for training, and 3D liver datasets are still very limited. Additionally, 3D images have complex structures. A solution to address data limitations and the complexity of 3D structures is to convert 3D images into 2D using Multi-Planar Reconstruction (MPR) techniques. The most parameterheavy parts of U-Net are the encoder and decoder. The large number of parameters can lead to overfitting. One way to reduce parameters is by using ResNet. The ResNet architecture has residual blocks equipped with skip connections that can accelerate the training process. Although the bridge does not contribute as many parameters as the encoder and decoder, it has high computational complexity because it is responsible for effectively linking information between the encoder and decoder. One CNN architecture that can enhance model performance is MobileNet. MobileNet uses depthwise separable convolutions, resulting in a lightweight and fast model. This study achieved an accuracy of 99.08%, sensitivity of 99.49%, specificity of 91.76%, IoU of 99.04%, and F1-Score of 99.52%. These results indicate that the model in this study effectively performs segmentation and identifies the liver organ as white and the background as black with high precision.

Item Type:	Thesis (Undergraduate)
Subjects:	Q Science > QA Mathematics > QA75-76.95 Calculating machines > QA76.9.D343 Data mining. Database searching. Big data.
Divisions:	08-Faculty of Mathematics and Natural Science > 44201-Mathematics (S1)
Depositing User:	Annisa Nurba Iffah'da
Date Deposited:	19 Aug 2024 03:57
Last Modified:	19 Aug 2024 03:57
URI:	http://repository.unsri.ac.id/id/eprint/154362

Actions (login required)

View Item