Detection and Tracking of Point Features

The factorization method described in this series of reports requires an algorithmto track the motion of features in an image stream. Given thesmall inter-frame displacement made possible by the factorization approach,the best tracking method turns out to be the one proposed by Lucas andKanade in 1981.The method defines the measure of match between fixed-size feature windowsin the past and current frame as the sum of squared intensity differences overthe windows. The displacement is...