Bias compensated least squares estimate of the center of rotation

It is shown that a recently published least squares method for the estimation of the average center of rotation is biased. Consequently, a correction term is proposed, and an iterative algorithm is derived for finding a bias compensated solution to the least squares problem. The accuracy of the proposed bias compensated least squares method is compared to the previously proposed least squares method by Monte-Carlo simulations. The tests show that the new method gives a substantial improvement in accuracy. (C) 2003 Elsevier Science Ltd. All rights reserved.