Towards Scalable Summarization of Consumer Videos Via Sparse Dictionary Selection

摘要: The rapid growth of consumer videos requires an effective and efficient content summarization method to provide a user-friendly way manage browse the huge amount video data. Compared with most previous methods that focus on sports news videos, personal is more challenging because its unconstrained lack any pre-imposed structures. We formulate as novel dictionary selection problem using sparsity consistency, where key frames selected such original can be best reconstructed from this representative dictionary. An global optimization algorithm introduced solve model convergence rates O(1/K2) (where K iteration counter), in contrast traditional sub-gradient descent O(1/√K). Our provides scalable solution for both frame extraction skim generation, one select arbitrary number represent videos. Experiments human labeled benchmark dataset comparisons state-of-the-art demonstrate advantages our algorithm.