Image metric-based correction (Autofocusing) of motion artifacts in high-resolution trabecular bone imaging

Purpose: To evaluate the performance of the autofocusing (AF) motion correction technique in high-resolution trabecular bone imaging where image signal-to noise ratio (SNR) is limited. Materials and Methods: Raw data from 26 clinical three-dimensional (3D) wrist exams were motion corrected using AF for both in-plane rotation and translation. Changes in image metrics (a measurement of image sharpness) and structural parameters subsequently, computed, were used to gauge the performance of the AF algorithm, and comparisons were made with translation-only navigator-corrected results. Results: On average, AF generated images with higher image sharpness compared to the navigator echo technique. The average normalized gradient squared (NGS) metric improved by 0.40%, 0.73%, and 0.84%, respectively, following translation-only navigator, translation-only AF and combined rotation/translation AF. For all structural parameters, the rotation/translation AF resulted in an approximately two-fold greater change compared to the navigator technique. Conclusion: The data provide evidence that errors from subtle translational and rotational motion in the structural parameters in high-resolution trabecular bone images are alleviated by AF and that the resulting improvements are superior to translation-only 2D navigator correction.