Recently, with the enormous growth of online videos, fast video retrievalresearch has received increasing attention. As an extension of image hashingtechniques, traditional video hashing methods mainly depend on hand-craftedfeatures and transform the real-valued features into binary hash codes. Asvideos provide far more diverse and complex visual information than images,extracting features from videos is much more challenging than that from images.Therefore, high-level semantic features to represent videos are needed ratherthan low-level hand-crafted methods. In this paper, a deep convolutional neuralnetwork is proposed to extract high-level semantic features and a binary hashfunction is then integrated into this framework to achieve an end-to-endoptimization. Particularly, our approach also combines triplet loss functionwhich preserves the relative similarity and difference of videos andclassification loss function as the optimization objective. Experiments havebeen performed on two public datasets and the results demonstrate thesuperiority of our proposed method compared with other state-of-the-art videoretrieval methods.
展开▼
