Key Terms: Motion Estimation, Optical Flow, Non-rigid Motion. Temporal Multi-scale Models for Flow and Acceleration

The quadrants' boundary intensity discontinuity a result of of the video-camera consisting of four separate A/D banks. As a result, ow estimation at the quadrant boundaries is inaccurate. The problem could be overcome by local gain compensation. The dense ow algorithms of Black and Anandan 6] is used. The algorithm's parameters were changed to achieve best measurement results. The same eeect could have been achieved by dividing the right side of Equation (3) by s for all scales prior to error summation. The book is manually segmented in the rst image and tracked automatically afterwords using our multi-scale parameterized ow model. ABSTRACT A model for computing image ow in image sequences containing a very wide range of instantaneous ows is proposed. This model integrates the spatio-temporal image derivatives from multiple temporal scales to provide both reliable and accurate instantaneous ow estimates. The integration employs robust regression and automatic scale weighting in a generalized brightness constancy framework. In addition to instantaneous ow estimation the model supports recovery of dense estimates of image acceleration and can be readily combined with parameterized ow and acceleration models. A demonstration of performance on image sequences of typical human actions taken with a high frame-rate camera is given.