Emperor Penguin optimized event recognition and summarization for cricket highlight generation

Cricket highlight generation is the process of summarizing a full-length video to a shortened form which should preserve the important moments present in the original video. In this paper, a new approach has been proposed for recognizing the key events and summarization. Audio features are initially used for extracting the excitement clips. Then, the important events like replay, players, umpires, spectators, and players gathering are extracted from each clip. Here, a hybrid deep neural network with Emperor Penguin optimization (HDNN-EPO) is proposed for labeling the excitement concepts presented in the cricket video based on the observed events automatically. These labeled concepts are then selected based on the importance degree and concatenated in temporal order to form highlights. The efficiency of the proposed method is proved through the experimental results and it outperforms the other existing approaches. Also, the extracted highlights have been compared with the manually-generated highlights by the sports television channel.