Video captioning based on dual learning via multiple reconstruction blocks

Open Webpage

Bahy Helmi Hartoyo Putra, Cheol Jeong

Published: 2024, Last Modified: 07 Nov 2025Image Vis. Comput. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Multi-stage regularization effectively enhances video captioning performance.•The number of blocks determines the trade-off between regularization and performance.•Inference can be fast as only the first captioning block is needed.