Jurnal Informatika
Vol 1, No 1 (2014): Jurnal INFORMATIKA

Analisa Size Citra Sel Tunggal Nukleus Menggunakan Global Threshold dan Operasi Kanal Warna

Yudi Ramdhani (AMIK BSI Bandung)
Dwiza Riana (Unknown)
Ade Mubarok (Unknown)



Article Info

Publish Date
11 Mar 2016

Abstract

ABSTRAK Pap Smear merupakan salah satu metode deteksi dini pencegahan kanker serviks. Sel mikroskopik Pap Smear dideteksi oleh ahli patologi dan dokter untuk menentukan sel tersebut normal atau abnormal. Data yang digunakan dalam penelitian ini menggunakan data citra Harlev, yang terdiri dari 280 citra yang sudah dikategorikan ke dalam 7 kelas sel normal yang meliputi Normal Superficial, Normal Intermediate, and Normal Columnar dan 4 kelas lainnya adalah kategori kelas citra sel abnormal yang meliputi Mild (Light) Dyplasia, Moderate Dysplasia, Severe Dysplasia dan Carcinoma In Situ. Pada analisa size citra sel tunggal nukleus, dilakukan segmentasi luas dan keliling. Metode yang digunakan untuk segmentasi luas nukleus yaitu operasi kanal warna RGB, metode Global Threshold dan deteksi tepi Canny. Sedangkan untuk segmentasi keliling menggunakan metode operasi kanal warna Grayscale, metode Global Threshold dan deteksi tepi Canny. Hasil penelitian memperlihatkan bahwa perhitungan luas menggunakan operasi kanal warna RGB lebih mendekati hasil perhitungan manual dari data Herlev. Nilai hasil korelasi dengan jangkauan nilai 0.547 dan nilai korelasi menggunakan metode Global Threshold dengan jangkauan nilai 0.107. Nilai korelasi perhitungan keliling menggunakan operasi kanal warna Grayscale lebih mendekati hasil perhitungan manual dari data Herlev. Hasil nilai korelasi perhitungan menggunakan operasi kanal warna Grayscale dengan jangkauan nilai 0.930, sedangkan hasil nilai korelasi perhitungan menggunakan metode Global Threshold dengan jangkauan nilai 0.892. Kata Kunci : Pap Smear, Nukleus, Analisa Size, Global Threshold, Operasi Kanal Warna ABSTRACT Pap smear is one of early detection method for prevent serviks cancer. Microscopic cell on pap smear detected by patology expert and doctor to determine is the cell normal or abnormal. The data used in this research is using Harlev image data, which consists of 280 images that have been categorized into 7 class of normal cell including Normal cells Superficial, Normal Intermediate, and Normal Columnar and the other 4 class is an image class category of abnormal cell including images Mild (Light) Dyplasia, Moderate Dysplasia, Severe Dysplasia and Carcinoma In Situ. On analize single cell image size of nucleus, performed a segmentation area and perimeter. Using RGB color channel operation method, Global Threshold method and canny edge detection for segmentation area. While for perimeter segmentation is using Grayscale color channel operation method, Global Threshold method and canny edge detection. Result of experimentation showing that RGB color channel operation methoded to manual counting from the Herlev data, if we compare it with the extensive calculations using the Global Threshold. Value of the correlation with 0.547 range values and the correlation values using the Global Threshold with 0.107 range values. On both of normal single cell, Normal Superficial (NS) and Normal Columnar (NC) using calculation of Global Threshold method, has a significant wide correlationwith an average correlation above 0.4. Value of correlation counting using the Grayscale color channel operation color is tend to the results of the manual calculation of the Herlev data, compared with calculations using the Global Threshold method. Value result of corellation counting, using the Grayscale color channel operation with 0.930 range values, while using the Global Threshold method exist on 0.892 range values. In the circumference calculating of the Normal Superficialclass (NS) and Normal Intermediate (NI) using the Global Threshold method is better compared with calculations using the Grayscale color channel operation. Keywords : Pap Smear, Nukleus, Analisis Size, Global Threshold, Color Channel Operation

Copyrights © 2014






Journal Info

Abbrev

ji

Publisher

Subject

Computer Science & IT

Description

Jurnal Informatika respects all researchers Technology and Information field as a part spirit of disseminating science resulting and community service that provides download journal articles for free, both nationally and internationally. The editorial welcomes innovative manuscripts from Technology ...