Pengembangan Deteksi Tumor Payudara melalui Segmentasi Citra Mammogram Berdasarkan Metode Active Contour Lankton Berbasis Gui Matlab

Maulida, Atika (2024) Pengembangan Deteksi Tumor Payudara melalui Segmentasi Citra Mammogram Berdasarkan Metode Active Contour Lankton Berbasis Gui Matlab. S2 thesis, Universitas Andalas.

	Text (Cover dan Abstrak) Cover dan Abstrak.pdf - Published Version Download (114kB)
	Text (bab 1) BAB 1.pdf - Published Version Download (113kB)
	Text (Bab 5) Bab 5.pdf - Published Version Download (36kB)
	Text (Daftar Pustaka) DAFTAR PUSTAKA.pdf - Published Version Download (180kB)
	Text (Tesis Full dan Turnitin) Tesis+turnitin.pdf - Published Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Breast tumors can be divided into benign tumors and malignant or cancerous tumors. Many risk factors are associated with the appearance of breast tumors, one of which is age. Early detection of breast tumors can be done through mammography procedures that will produce mammogram images. This research develops a MATLAB GUI-based program using the Active Contour Lankton method for segmentation and detection of tumor diameter in mammogram images. Before segmentation, the image undergoes an enhancement process with Intensity Adjustment method to reduce noise and clarify the suspected tumor lesion area. This program focuses on segmentation and diameter measurement of suspicious lesions, which is the first step in the diagnosis process. In this study, a total of 15 mammogram image data from 15 different patients taken from Dr. M. Djamil Hospital Padang were used. The results showed that the program was able to perform segmentation and measurement of tumor lesion diameter with good sensitivity. The program successfully detected 8 patients with tumors that have been confirmed by doctors and 4 normal patients. However, there were 3 false positives, where the program detected tumors in patients who were actually normal. This shows the need to improve accuracy, especially in reducing false positives. Therefore, it is recommended to conduct further development on this program to improve the segmentation accuracy and reduce detection errors, so that this program can be relied upon in supporting clinical diagnosis and early detection of breast tumors.

Actions (login required)

View Item