Parameter estimation-based discrimination method for osteoporosis stage with ultrasound bone densitometer.

This paper proposes a new method for predicting the stage of osteoporosis by estimating bone-density parameters using an ultrasonic-bone densitometer. Using the developed ultrasonic bone densitometer, the ultrasonic signal passing through the phalanx of the finger was measured, and the measured signal was preprocessed to generate input data for parameter estimation. The Nakagami probability-density function was used for modeling this data, and optimal bone-density parameters were estimated using an optimization problem - to predict the stage of osteoporosis. The proposed method was verified through in vitro and in vivo experiments. In phantom experiments, cubes with different materials (ABS plastic, PLA plastic, Nylon) were used, and each cube had a different density through internal density so that it could mimic bones. The experiments were conducted using them and the results were compared. The experimental results confirmed that there was a correlation between the density and estimated parameters for all three materials. In the in vivo experiment, the osteoporosis stages were classified as normal, osteopenia, and osteoporosis, and the bone-density parameters were estimated for the participants of each group. Thus, a basis for predicting the stage of osteoporosis was created, and data from new participants were collected and verified as test data. As a result, normal and abnormal groups (osteopenia and osteoporosis) were predicted.