Deep multi-scale pyramidal features network for supervised video summarization

Open Webpage

Habib Khan, Tanveer Hussain, Samee Ullah Khan, Zulfiqar Ahmad Khan, Sung Wook Baik

Published: 01 Jan 2024, Last Modified: 24 Jul 2025Expert Syst. Appl. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•DPT extracts deep, specific, and global features from video frames.•Instead of considering a single layer, features are extracted from multi-scales.•PRB is designed that boost the frames representations before passing to CH.•Progressive features fusion ensures features refinement before predictions.