Enhancing Osteoporosis Detection with Hybrid Fuzzy Logic Preprocessed Convolutional Neural Networks

Woud Majid Abed^{1, 2,*}, Murtadha M. Hamad², Azmi Tawfeq Hussein Alrawi²

¹Department of Basic Sciences, College of Dentistry, University of Baghdad, Baghdad, Iraq

²Collage of Computer Science and Information Technology, University of Anbar, Rmadi, Iraq

Emails: wou22c1001@uoanbar.edu.iq; dr.mortadha61@uoanbar.edu.iq; azmi.alrawi@uoanbar.edu.iq;

Abstract

This paper applies deep learning techniques in classifying X-ray images to detect osteoporosis. Osteoporosis, a bone weakness condition, increases the risk of fractures; therefore, accurate early diagnosis is essential in management. We have designed a hybrid method called Fuzzy Logic Preprocessed Convolutional Neural Network, or FLPCNN, wherein fuzzy logic is used at the preprocessing step to handle uncertainty and imprecision of features extracted from X-ray images. This paper used a dataset of X-ray images, and the FLPCNN model was applied, classifying them into osteoporotic and non-osteoporotic with quite an accuracy of 100%. Fuzzy logic preprocessing combined with Convolutional Neural Networks (CNN) enhances the model’s classification accuracy and interpretable decisions. The proposed method would be a new way to cut down diagnostic errors and improve patient outcomes, opening ways for further research into deep learning techniques applied in healthcare.

Keywords: Osteoporosis Detection; X-ray Classification; Deep Learning; Convolutional Neural Networks, Fuzzy Logic