On the Use of Deep Learning for Video Classification

The video classification task has gained significant success in the recent years. Specifically, topic more attention after emergence of deep learning models as a successful tool for automatically classifying videos. In recognition importance and to summarize this task, paper presents very comprehensive concise review on topic. There are several existing reviews survey papers related scientific literature. However, do not include state-of-art works, they also have some limitations. To provide an updated review, highlights key findings based models. discussed way future research directions. This mainly focuses type network architecture used, evaluation criteria measure success, datasets used. make self-contained, methods towards automatic well explained summarized. Moreover, clear insight newly developed architectures traditional approaches is provided. critical challenges benchmarks highlighted evaluating technical progress these methods. summarizes benchmark performance matrices classification. Based compact, complete, proposes new directions solve challenging problem.