Nollywood Movie Sequences Summarization Dataset for Machine Learning Applications

Summarization in recent times has become one of the most exploited areas of natural language processing tasks, due to significant increase in data volume, which is more than terabytes and petabytes. In this article, we present a custom dataset for Nollywood movie sequences, an extensive collection of 13 Nollywood movies representing various genres, including drama, comedy and action. This custom dataset was impendent into pre-train benchmark dataset TVSum following the same file format, frame size and others in order to suit our purpose. Fast Forward Moving Picture Expert Group (FFMPEG) was used for video encoding during pre-processing and final skimming. The proposed framework is effective for creating movie summaries with a high user satisfaction rate and having reasonably low computing requirements.