AI Paper Atlas

Research map / Video / Video-language understanding

Reconstructive Sequence-Graph Network for Video Summarization

Bin Zhao, Haopeng Li, Xiaoqiang Lu, Xuelong Li

TPAMI 2021 · Video › Video-language understanding · approach: Encoder-decoder video captioning · 131 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 graph-based shot encoder instead of sequence encoder (Enc(S) -> Enc(G)) (ablation)RSGN without reconstructor · SumMeF-score: 0.397 -> 0.419Improved
add graph-based shot encoder instead of sequence encoder (Enc(S) -> Enc(G)) (ablation)RSGN without reconstructor · TVsumF-score: 0.556 -> 0.564Improved
replace second-level sequence encoder with graph encoder (Enc(S,S) -> Enc(S,G)) (ablation)RSGN without reconstructor · SumMeF-score: 0.416 -> 0.428Improved
replace second-level sequence encoder with graph encoder (Enc(S,S) -> Enc(S,G)) (ablation)RSGN without reconstructor · TVsumF-score: 0.583 -> 0.587Improved
use hierarchical encoder instead of plain graph encoder (Enc(G) -> Enc(S,G)) (ablation)RSGN without reconstructor · SumMeF-score: 0.419 -> 0.428Improved
add content reconstructor Rec(C) (ablation)Enc(S,G) · TVsumF-score: 0.587 -> 0.597Improved
add dependency reconstructor Rec(D) (ablation)Enc(S,G) · TVsumF-score: 0.587 -> 0.599Improved
add both reconstructors, full RSGN, over Enc(S,G) without reconstructor (ablation)Enc(S,G) · SumMeF-score: 0.428 -> 0.450Improved
add supervision to RSGN (RSGN_uns -> RSGN) (ablation)RSGN · SumMeF-score: 0.423 -> 0.450Improved
edge weight by dot product vs Gaussian (ablation)graph encoder · SumMeF-score: 0.413 -> 0.419No clear effect
edge weight by dot product vs concatenation (ablation)graph encoder · TVsumF-score: 0.566 -> 0.564No clear effect
add reconstructors to graph-only encoder on VTW (ablation)RSGN Enc(G) · VTWF-score: 0.450 -> 0.468Improved

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