Video Summarization Using Machine Learning Techniques: An Overview

Technological advancement is a persistent aspect in our lives, and it emerges and grows at an exponential rate. Every day, a large quantity of data is generated in the form of text, image, audio and video. To process large amounts of video data, like movies, social media data and Surveillance data, requires a large amount of storage. Therefore, minimizing these videos by processing only vital content, takes a long processing time. To extract the key content from the video, which is a time-consuming procedure for the viewer, the entire video must be watched to overcome such challenges, video summary can be used to deal with and process lengthy videos. This paper discusses the various machine learning strategies used for summarising videos. Also, this paper presents video processing approaches such as key framing and skimming used for summarization in dynamic environments such as surveillance. Further, this study emphasizes the primary uses of video summarising in both dynamic and static environment. Furthermore, it deals with the datasets used for video summarisation and helps a clear understanding on various techniques applied for video summarisation at both static and dynamic environments.