An algorithm for key-frame determination in digital video

An algorithm for determining the key-frames on a digital video is presented in this research paper. The algorithm is based on the idea of stochastic invariance, defined in this paper. The invariant space for the video sequence between key-frames is processed only once and compressed only once. Part of the processing of the invariance space pertains to the noise reduction via ensemble averaging of the invariant pixels over the group of frames which are invariant. The other part pertains to the compression of the ensemble averaged pixels.