Ultrasonic image processing for tendon injury evaluation

Abstract

Image texture analysis techniques have been applied for evaluating tendon injun es in racehorses. Ultrasound images from superficial and deep1 digital flexor tendons have been characterized and a measure for quantifying injury grade has been proposed.Wivelet-based texture segmentation and classification algorithms have been developed for quantification of the injury grades. Discrete wavelet packet frames and standard discrete wavelet transforms are used to measure the percentage of the tendon injury and tendon fiber alignment, respectively. Results demonstrating the effectiveness of the approach in predicting the percentage of the tendon injury as well as the degree of tendon fiber alignment are presented.