Klasifikasi Tiga Jenis Kanker Kulit Menggunakan Metode Fraktal Dengan Klasifikasi K-Nearest Neighbour

Muhammad, Zukhruf Abidi (2022) Klasifikasi Tiga Jenis Kanker Kulit Menggunakan Metode Fraktal Dengan Klasifikasi K-Nearest Neighbour. Undergraduate Thesis thesis, Institut Teknologi Telkom Purwokerto.

	Text COVER.pdf Download (284kB)
	Text ABSTRACK.pdf Download (7kB)
	Text ABSTRAK.pdf Download (134kB)
	Text BAB 1.pdf Download (22kB)
	Text BAB 2.pdf Download (543kB)
	Text BAB 3.pdf Download (340kB)
	Text BAB 4.pdf Restricted to Registered users only Download (264kB) | Request a copy
	Text BAB 5.pdf Download (14kB)
	Text DAFTAR PUSTAKA.pdf Download (237kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (1MB) | Request a copy

Abstract

Examinations to detect and classify the severity of skin cancer patients are currently carried out by trained medical personnel manually and require a long time, so that the right technology is needed to detect skin cancer early so that it can be treated immediately. With increasingly developed technology, a detection system for skin cancer can be made with image segmentation using fractal methods and KNearest Neighbor classification. Image data will be processed using the fractal method to find the unique characteristics of each image data. Then the K-NN classification will calculate the closest distance between the test data and the training data, the K-NN work process starts from determining the k parameter (the number of closest neighbors) then calculating the square of the distance of each object to the sample data. The classification will be divided into three classes, namely: melanoma skin cancer, nevus skin cancer, and seborrheic keratosis skin cancer. In this study, the highest accuracy was 90.7% with a computation time of 93.2 seconds by using a 512 x 512 image resize and the value of K = 1 with training image data of 300 and test image of 150, the value of K = 1 is the most optimal K value among K-NN Classification. Keyword: Skin cancer, Matlab, Fraktal, K-Nearest Neighbor (K-NN).

Actions (login required)

View Item