AI Paper Atlas

Research map / Video / Video segmentation and tracking

Adaptive Video Highlight Detection by Learning from User History

Mrigank Rochan, Mahesh Kumar Krishna Reddy, Linwei Ye, Yang Wang

ECCV 2020 · Video › Video segmentation and tracking · approach: Self-supervised video representation learning · 35 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add T-AIN layers conditioned on user historyH-FCSN · PHD-GIFs testmAP (%): 15.04 -> 16.73Improved
combine user history by aggregation (add averaged history to frame features)H-FCSN · PHD-GIFs testmAP (%): 15.04 -> 15.73Improved
predict affine parameters from user history instead of learning them directly (ablation)H-FCSN · PHD-GIFs test (affine parameter ablation)mAP (%): 15.04 -> 16.73Improved
learn affine parameters instead of fixing them to 1 and 0 (ablation)H-FCSN · PHD-GIFs test (affine parameter ablation)mAP (%): 14.64 -> 15.04Improved
learn affine parameters instead of fixing them (ablation)H-FCSN-aggregate · PHD-GIFs test (affine parameter ablation)mAP (%): 14.87 -> 15.73Improved
add self-attention to history encoder (ablation)Adaptive-H-FCSN · PHD-GIFs testmAP (%): 16.73 -> 16.37Hurt
increase user history from 1 element to full history (ablation)Adaptive-H-FCSN · PHD-GIFs test (history size)mAP (%): 15.57 -> 16.73Improved
increase user history from 1 element to full history (ablation)H-FCSN-aggregate · PHD-GIFs test (history size)mAP (%): 15.62 -> 15.73No clear effect
use history of one element vs none (ablation)Adaptive-H-FCSN · PHD-GIFs test (history size)mAP (%): 15.04 -> 15.57Improved
pretrain on large-scale highlight detection (PHD-GIFs) for summarizationH-FCSN · SumMe (20% test, no SumMe training)F-score (%): 43.3 -> 44.4Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy