Social-Media Video Summarization Using Convolutional Neural Network and Kohnen’s Self Organizing Map

Video Summarisation (VS) techniques provide concise representations of original videos. This research presents a novel approach to extracting keyframes based on Convolutional Neural Network and Self-Organizing Map (SOM). The technique analyses videos frame by frame. Convolutional Neural Network is used to extract the deep-level feature vectors. Using Self-Organizing Map clustering, the feature descriptors corresponding to the frames are divided into keyframes and non-key frames. This SOM-based video summarization method effectively selects the most representative frames from the extracted feature vector. The proposed approach produces better results than other state-of-the-art video summarising techniques. In the comprehensive experimental analysis of two benchmark datasets, our methodology produced average F-scores of 0.81 and 0.82, respectively. The outcomes demonstrate that the method consistently generates high-quality summaries for videos across various categories.