KOMBINASI ARSITEKTUR VISUAL GEOMETRY GROUP, BOTTLENECK LAYER, GATE RECURRENT UNITS, DAN ATTENTION GATE PADA KLASIFIKASI PENYAKIT KANKER SERVIKS MENGGUNAKAN CITRA PAP SMEAR

BAGAS, BAGAS and Desiani, Anita and Suprihatin, Bambang (2025) KOMBINASI ARSITEKTUR VISUAL GEOMETRY GROUP, BOTTLENECK LAYER, GATE RECURRENT UNITS, DAN ATTENTION GATE PADA KLASIFIKASI PENYAKIT KANKER SERVIKS MENGGUNAKAN CITRA PAP SMEAR. Undergraduate thesis, Sriwijaya University.

	Image RAMA_44201_08011282126063_cover.jpg - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (123kB)
	Text RAMA_44201_08011282126063.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_44201_08011282126063_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (9MB) | Request a copy
	Text RAMA_44201_08011282126063_0011127702_0026017102_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (606kB)
	Text RAMA_44201_08011282126063_0011127702_0026017102_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_44201_08011282126063_0011127702_0026017102_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (867kB) | Request a copy
	Text RAMA_44201_08011282126063_0011127702_0026017102_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (5MB) | Request a copy
	Text RAMA_44201_08011282126063_0011127702_0026017102_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (244kB) | Request a copy
	Text RAMA_44201_08011282126063_0011127702_0026017102_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy

Abstract

Superficial-intermediate, parabasal, koilocytotic, dyskeratotic, and metaplastic are types of cells in the cervix that are at risk of causing cancer. Cervical cancer can be detected early using several methods, one of which is the Pap smear. Early detection of this disease can be achieved through the use of deep learning methods. This study proposes a model combining the Visual Geometry Group Bottleneck Layer architecture, Gate Recurrent Units, and Attention Gate for cervical cancer classification using Pap smear images. In the first block, the VGG architecture is used to capture detailed features in complex features. In the 14th block, before classification, a Bottleneck Layer is added to reduce the number of parameters and prevent overfitting before entering the classification layer. Gate Recurrent Units are added after the Bottleneck Layer to reuse features that were discarded due to feature dimension reduction from the Bottleneck Layer, and an Attention Gate is inserted to focus the model on relevant features while filtering out irrelevant ones. In the last block, the model uses a fully connected layer and a softmax activation function to provide prediction results. The proposed model achieved an average accuracy of 99.6%, sensitivity of 95.4%, specificity of 99.7%, F1-score of 95.4%, and Cohen's kappa of 95.2%, indicating that the model is capable of classifying Pap smear images effectively. In this study, the proposed model performed very well in recognizing the superficial-intermediate class compared to the parabasal, koilocytotic, dyskeratotic, and metaplastic classes, with a sensitivity of 99.8%, indicating that the model is very good at recognizing images from the superficial-intermediate class. Although the values obtained for the parabasal, koilocytotic, dyskeratotic, and metaplastic classes are still below those of the superficial-intermediate class, the results obtained are already very good because the sensitivity values for the parabasal, koilocytotic, dyskeratotic, and metaplastic classes are 90%.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Kanker Serviks, Klasifikasi, VGG, Bottleneck layer, GRU
Subjects:	Q Science > QA Mathematics > QA299.6-433 Analysis > Q334.A755 Artificial intelligence. Computational linguistics. Computer science. Q Science > QA Mathematics > QA8.9-QA10.3 Computer science. Artificial intelligence. Computational complexity. Data structures (Computer scienc. Mathematical Logic and Formal Languages
Divisions:	08-Faculty of Mathematics and Natural Science > 44201-Mathematics (S1)
Depositing User:	Bagas Bagas
Date Deposited:	19 Sep 2025 10:05
Last Modified:	19 Sep 2025 10:05
URI:	http://repository.unsri.ac.id/id/eprint/184102

Actions (login required)

View Item