Authors
Abstract
Background: The burst fractures of the vertebrae include about 15% of spinal column injuries, and the most common location is in the lumbar back area. The diagnosis of fracture types and associated injuries is usually done by using simple radiography, CT scan, and MRI. In this research, we tried to present a new method for fracture detection with the thermographic images.
Methods: The present study was a preliminary study, which was conducted on a set of thermal images obtained from a Clinical Center in California. The diagnosis (detection) of unstable burst fractures of the lumbar spinal column was performed based on the thermal pattern using the Fuzzy C-Means (FCM) clustering method and recursive connected components algorithm. In this study, the procedure was validated and confirmed by examinations and evaluation previously made by an orthopedic surgeon on the same patients
Results: After applying the preprocessing steps and FCM clustering on the image, the clusters belonging to the lumbar spine, which center was at the first and second places of the clusters centers matrix, were gathered together. Then, the unstable lumbar spinal burst fractures were diagnosed based on the components labeling technique. From the 130 thermographic images, 93 showed fracture and in 33 no fracture was seen. This confers with the CT scan images, and shows 95% accuracy.
Conclusion: The method presented in this article is a non-invasive and cost-effective approach for the diagnosis of unstable burst fractures of lumbar spinal column. The techniques and tips derived empirically, based on the scientific principles of this research, can help the physicians to quickly diagnose the burst fractures of lumbar spinal column based on the analysis of thermal images.
Keywords
)