Predicting finger joint angle using ultrasound images of the forearm

Abstract

Identifying the pose of fingers is used for basic research on the human hand and analyzing dexterous movements. Joint angles can be extracted from instrumented gloves or by tracking markers with video cameras. Such methods are cumbersome, however, and can even impede complex manipulatory tasks as in surgery. We propose a novel solution, using ultrasound images of the muscles in the forearm to determine finger pose. We calibrate its use for monitoring the contraction and relaxation of the flexor digitorum superficialis muscles in the index finger, relative to marker-based tracking, and demonstrate machine classification of finger angle from ultrasound.