A Survey on Evolution of Video Summarization Techniques in the Era of Deep Learning

By choosing the most instructive segments of the video content, video summarization systems seek to produce a succinct and comprehensive summarization. The most advanced techniques are those that use contemporary Deep Neural Network (DNN) architectures. Several strategies were developed in the past. The most current developments in the field are highlighted while providing a full analysis of current deep-learning-based approaches for broad video summarization. This study reviews the video summary technologies and an examination of key elements of deep-learning-based algorithms. An analysis of deep learning-based algorithms is provided to analyze the development of deep learning-based algorithms for video summarization and to offer suggestions for the pace of future research. The results of these comparisons, as well as some previously discussed challenges regarding the datasets and the applicability of evaluation procedures, will leverage future directions.