Segmentasi Radiografi Tangan Pasien Artritis Rematoid dengan Pendekatan Branches Filtering

Andrijani Sumarahinsih (Universitas Merdeka Malang)
Handono Kalim (Universitas Brawijaya Malang)
Yuyun Yueniwati (Universitas Brawijaya Malang)
Agus Naba (Universitas Brawijaya Malang)

JASIEK (Jurnal Aplikasi Sains, Informasi, Elektronika dan Komputer)
Vol 4, No 2 (2022): DESEMBER 2022

This research is for the possibility of automating the assessment of hand radiographic joint damage in rheumatoid arthritis. The objective of this research is to design a segmentation algorithm to obtain the area which is the joint space object. This study has collected 46 radiographic images of the hands. Image preprocessing is performed using adaptive thresholds and the concept of morphological gradient, then segmentation is performed using branches filtering approach. Analysis based on accuracy, sensitivity and specificity compared with manual segmentation by experts. The results of segmentation research preceded by preprocessing produce better images with an accuracy of 96.413%, a sensitivity of 72.121% and a specificity of 97.891% in the highest percentage.