Summarizing Videos with Attention

5 Dec 2018Jiri FajtlHajar Sadeghi SokehVasileios ArgyriouDorothy MonekossoPaolo Remagnino

In this work we propose a novel method for supervised, keyshots based video summarization by applying a conceptually simple and computationally efficient soft, self-attention mechanism. Current state of the art methods leverage bi-directional recurrent networks such as BiLSTM combined with attention... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Video Summarization SumMe VASNet F1-score (Canonical) 49.71 # 1
F1-score (Augmented) 51.09 # 1
Video Summarization TvSum VASNet F1-score (Canonical) 61.42 # 1
F1-score (Augmented) 62.37 # 1