Diagnostic of Osteoporosis Using Backpropagation Neural Networks
In this study, an artificial neural network (ANN) using backpropagation was utilized to categorize bone images into either healthy or osteoporotic categories based on various statistical operations. An input matrix was constructed containing the six statistical features of 125 samples, representing...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
middle technical university
2025-06-01
|
| Series: | Journal of Techniques |
| Subjects: | |
| Online Access: | https://journal.mtu.edu.iq/index.php/MTU/article/view/2597 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In this study, an artificial neural network (ANN) using backpropagation was utilized to categorize bone images into either healthy or osteoporotic categories based on various statistical operations. An input matrix was constructed containing the six statistical features of 125 samples, representing X-ray images of knee joints, with 25 healthy and 100 osteoporotic samples. Of these, 70% were used for training, 15% for validation, and 15% for network testing. The classification efficiency of the neural network for the 125 samples was 97%. The research included analysis of arithmetic mean, standard deviation, variance, energy, homogeneity, and entropy values for the healthy bone samples. The backpropagation neural network (BNN) was trained with six inputs (representing the six statistical features), 80 hidden layers, and five outputs (two for healthy and three for osteoporotic conditions). A comparison of K-Nearest Neighbors (KNN), Logistic Regression, and BNN techniques applied to 2,350 images revealed that BNNs achieved the highest accuracy. This network has the potential to assist healthcare providers in both detecting the early stages of osteoporosis and developing appropriate treatment plans.
|
|---|---|
| ISSN: | 1818-653X 2708-8383 |