Signal processing techniques for accurate screening of wrist fractures

The common way for doctors to differentiate wrist injury in to a sprain or a fracture, is to take radiographs (X-ray), which expose patients to radiation. The purpose of this study is to explore a non-invasive method to screen for potential fractures. A small, computer run, hand-held system has been developed which consists of a vibration induction mechanism and a piezoelectric sensor for capturing the vibration signals. Two analyzing techniques were considered. The first involves extraction of wavelet coefficients from decomposition of data and the second applies Fast Fourier Transform to the data. Results of both techniques were then cluster analyzed to partition between fracture and sprain. The data were acquired from both the injured and uninjured wrists of six adult patients. This study is currently being evaluated on children’s wrists.