Video Stabilization using KLT features and Firefly Optimisation

We have proposed a novel video stabilization method is based on reducing the features points in beizer curve. KLT feature tracking method is used for features tracking and initialized by SURF features of first reference frame. In subsequent frames these corresponding features are tracked and by image warping method followed by firefly tuned beizer curve smoothing, image is transformed to stable state. MATLAB’s computer vision toolbox saved our time to write the code of SURF features, KLT tracking, image warping etc and we could focus on beizer curve smoothing algorithm.