Explainable AI-Driven Intelligent Video Summarization for Next-Generation Video Production and Content Management

Abstract

This chapter reviews the development of video summarisation using handcrafted features and deep learning models, the recent advances in transformer-based, vision-language, multimodal learning, and large language models for video summarisation. It introduces the main steps of the summarization pipeline: keyframes extraction, keyshots generation, importance score prediction and summary generation. Particular focus is given to explainability techniques like LIME, Grad-CAM, attention maps and multi-granular explanations, and how frame, shot, semantic, and story-level explanations can help to increase the transparency and user trust. Applications in film and media production, OTT platforms, news, education and surveillance are also discussed as well as challenges concerning explainability evaluation, computational complexity, privacy, ethics and scalability. The future involves explainable LLMs, personalized summarization, human in the loop AI, interactive systems and foundation models.