A thinning algorithm for equine tendon structure identification from 2D ultrasound images

In this paper, we propose a new shock filter algorithm for the thinning of hyperechoic structures observed on ultrasound (US) images of the equine superficial digital flexor tendon (SDFT). Implementation of the new algorithm is presented and applied on in vivo US data sets of a healthy and of an injured SDFT scanned with 7.5 MHz linear array transducer (SSD-2000-7.5, Aloka). Quantitative and qualitative results on the enhanced images show a coherent distribution of SDFT fiber bundles. In the case of normal SDFTs, the calculated number and area were respectively 48plusmn13 and 1.39plusmn0.48 mm . In the case of injured SDFTs, the calculated number and area were respectively 40plusmn9 and 1.80plusmn0.68 mm². The 3D reconstruction of the SDFT allowed a better assessment of the fiber bundles alignment along the SDFT loading axis, as well as the injured areas. In conclusion, segmentation results demonstrated the potential of the new thinning algorithm on the SDFT structure characterization.