Optimasi Citra Mammogram Menggunakan Teknik CLAHE untuk Klasifikasi Tumor Payudara dengan Arsitektur ResNet50

Saraswati, Adinda Putri Budi (2025) Optimasi Citra Mammogram Menggunakan Teknik CLAHE untuk Klasifikasi Tumor Payudara dengan Arsitektur ResNet50. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 21081010201.-cover.pdf Download (4MB)
	Text (Bab 1) 21081010201.-bab1.pdf Download (25kB)
	Text (Bab 2) 21081010201.-bab2.pdf Restricted to Repository staff only until 5 December 2027. Download (686kB)
	Text (Bab 3) 21081010201.-bab3.pdf Restricted to Repository staff only until 5 December 2027. Download (684kB)
	Text (Bab 4) 21081010201.-bab4.pdf Restricted to Repository staff only until 5 December 2027. Download (767kB)
	Text (Bab 5) 21081010201.-bab5.pdf Download (11kB)
	Text (Daftar Pustaka) 21081010201.-daftarpustaka.pdf Download (150kB)
	Text (Lampiran) 21081010201.-lampiran.pdf Restricted to Repository staff only Download (155kB)

Abstract

Breast tumors represent a pathological condition characterized by uncontrolled and abnormal cell growth within breast tissue, which may be either benign or malignant. Among malignant cases, breast cancer has emerged as the leading cause of cancer-related mortality in women worldwide, with a steadily rising incidence year after year. In Indonesia, according to Globocan 2020 data, more than 66,000 new cases of breast tumors in women were recorded, resulting in approximately 22,000 deaths. These alarming figures highlight the critical need for accurate tumor characterization, as early and precise classification directly influences treatment decisions and patient outcomes. Consequently, there is an urgent demand for advanced technological solutions capable of delivering faster and more reliable diagnostic performance in breast tumor classification. This study aims to develop a deep learning-based classification model for breast tumors by employing transfer learning with the ResNet50 architecture applied to mammography images, categorizing them into three classes: normal, benign tumor, and malignant tumor. The research utilizes the mini-MIAS (Mammographic Image Analysis Society) dataset, which has been preprocessed using Contrast Limited Adaptive Histogram Equalization (CLAHE) to enhance local contrast and better reveal critical image features such as edges and texture. Additionally, extensive data augmentation and transformation techniques were applied to increase data diversity by introducing variations in breast position, orientation, and scale, thereby improving the model’s robustness and generalization ability. The optimally configured model achieved an accuracy of 0.9440, an AUC of 0.9765, and an F1-score of 0.94. These results demonstrate that the ResNet50 architecture, when combined with transfer learning, is highly effective for the task of multi-class breast tumor classification using mammographic images.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Sari, Anggraini Puspita	NIDN0716088605	anggraini.puspita.if@upnjatim.ac.id
Thesis advisor	Nurlaili, Afina Lina	NIDN0013129303	afina.lina.if@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA76.6 Computer Programming
T Technology > T Technology (General)

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Adinda Putri Budi Saraswati

Date Deposited:

05 Dec 2025 09:05

Last Modified:

05 Dec 2025 09:09

URI:

https://repository.upnjatim.ac.id/id/eprint/48064