Sistem Deteksi Penyakit Pengeroposan Tulang Dengan Metode Jaringan Syaraf Tiruan Backpropagation Dan Representasi Ciri Dalam Ruang Eigen

Authors

  • Is Mardianto Jurusan Teknik Informatika, Fakultas Teknologi Industri, Universitas Trisakti, Jln. Kyai Tapa No.1, Grogol, Jakarta Barat 11440
  • Dian Pratiwi Jurusan Teknik Informatika, Fakultas Teknologi Industri, Universitas Trisakti, Jln. Kyai Tapa No.1, Grogol, Jakarta Barat 11440

DOI:

https://doi.org/10.21512/commit.v2i1.494

Abstract

There are various ways to detect osteoporosis disease (bone loss). One of them is by observing the osteoporosis
image through rontgen picture or X-ray. Then, it is analyzed manually by Rheumatology experts. Article present the creation
of a system which could detect osteoporosis disease on human, by implementing the Rheumatology principles. The main areas
identified were between wrist and hand fingers. The working system in this software included 3 important processing, which
were process of basic image processing, pixel reduction process, pixel reduction, and artificial neural networks. Initially, the
color of digital X-ray image (30 x 30 pixels) was converted from RGB to grayscale. Then, it was threshold and its gray level
value was taken. These values then were normalized to an interval [0.1, 0.9], then reduced using a PCA (Principal Component
Analysis) method. The results were used as input on the process of Backpropagation artificial neural networks to detect the
disease analysis of X-ray being inputted. It can be concluded that from the testing result, with a learning rate of 0.7 and
momentum of 0.4, this system had a success rate of 73 to 100 percent for the non-learning data testing, and 100 percent for
learning data.
Keywords: osteoporosis, image processing, PCA, artificial neural networks
Dimensions

Plum Analytics

Downloads

Published

2008-05-31
Abstract 647  .
PDF downloaded 2305  .