BONE TUMOR SEGMENTATION FROM CT IMAGES


Creative Commons License

Catal Reis H., Bayram B.

SIGMA JOURNAL OF ENGINEERING AND NATURAL SCIENCES-SIGMA MUHENDISLIK VE FEN BILIMLERI DERGISI, cilt.7, ss.173-180, 2016 (Hakemli Dergi) identifier identifier identifier

Özet

Computer Aided Diagnosis (CAD) is one of the important topics of radiology and medical image processing. CAD systems brings second opinion to physicians by using algorithms and related software for diagnosis and treatments. In this study, an automatic and semi-automatic bone tumor extraction methods have been proposed. Bone tumors have been extracted from CT images. Proposed study consists of five steps which are; (i) pre-processing, (ii) segmentation, (iii) morphologic filtering, (iv) 3D modelling, (v) analysis and interpretation. MATLAB platform and 3D-Doctor software have been used for 2D and 3D processing respectively. This study has been realized by using retrospective nine CT data (male-female). Obtained results have been validated with clinical findings of radiology and orthopedic experts. Accuracy assessment of the study has been done by ROC curve.