Development and Evaluation of a Landmark-Based Method to Estimate Displacements and Strains in Soft Tissues Imaged During an Impact Using Ultrafast Ultrasound

摘要

Ultrafast ultrasound imaging allows observing the internal response of soft tissues subjected to impacts. However, the limited image quality and the large deformations associated with such loading make the generation of displacement and strain maps difficult. This study describes a new automatic method to estimate the 2D displacements and strains based on an existing tracking algorithm combined with filtering, regression and integration steps. The predictions were assessed using four test cases: (1) Predicted displacements were compared with expert tracking from past tests, (2) predicted displacements and strains were compared with the known field used to deform numerically an image, (3) internal strains predicted based on ultrafast ultrasound were compared with strains provided by digital image correlation on nearby external surfaces, and (4) for almost noise-free images that contain random patterns, predicted strains were compared with results provided by a commercial digital image correlation software package. The results are promising, in particular for 2D displacements and, to some extent, for filtered strain. This suggests that the method could be useful to interpret ultrafast ultrasound data collected for application in impact biomechanics.