Article Per Year (5 Year)
p-Index From 2019 - 2024
0
P-Index
This Author published in this journals
All Journal Jurnal Nasional Teknik Elektro dan Teknologi Informasi
Annisaa Sri Indrawanti
Institut Teknologi Adhi Tama
Computer Science & IT Control & Systems Engineering Electrical & Electronics Engineering Energy Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Deteksi Limfoblas pada Citra Sel Darah Menggunakan Fitur Geometri dan Local Binary Pattern Annisaa Sri Indrawanti; Eka Prakarsa Mandyartha
Jurnal Nasional Teknik Elektro dan Teknologi Informasi Vol 7 No 4: November 2018
Publisher : Departemen Teknik Elektro dan Teknologi Informasi, Fakultas Teknik, Universitas Gadjah Mada

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1098.356 KB)

Abstract

Lymphoblasts are white blood cell types of lymphocytes, which can mark leukemia. To identify lymphoblasts, an analysis of white blood cells is required. In this study, a computer-based automated system was proposed using digital image processing techniques to detect lymphoblasts by analyzing microscopic images of blood cells. This proposed method segments the components of white blood cells, which are cytoplasm and nucleus, using a new approach based on adaptive local thresholding techniques. After each cell component was segmented, the geometry features and texture were extracted. The texture feature used a local binary pattern (LBP) descriptor from the nucleus. The set of features was used to train the support vector machine classification algorithm in detecting lymphoblasts. The proposed method is able to segment correctly 264 of 269 total white blood cells, with 98.14% accuracy, out of 35 acute lymphoblastic leukemia images taken with the same camera with the same lighting conditions. The use of geometry features with 16 dimensional feature vector and LBP features with 256 dimensional feature vector result in accuracy of lymphoblast identification of 88.79% and 89.72% respectively. Better performance is obtained by combining two features, the geometry and the LBP with 272 dimensional feature vector, with classification accuracy of 94.32%.
Co-Authors Mandyartha, Eka Prakarsa