A Comparative Analysis of Query-Constrained Video Summarization Techniques

This paper is an analysis of the literature available in the field of query-constrained video summarization. For the sake of concision, five of the more recent, prominent research works have been explored. This research work will dive deep into the datasets used by each of the models, concerning their preparation methods, scale, description, and comparison. Next, the methodologies associated with each of these papers covering their models used, metrics employed, and draw parallels between them are described. The ablation analyses among model variants are discussed to gauge the importance of specific features in each paper. Finally, a description of each model's metric results has been provided for a quantitative comparison among the models, built on some standard metrics described by each of these papers.