AI Paper Atlas

Research map / Video / Video generation

SummDiff: Generative Modeling of Video Summarization with Diffusion

Kwanseok Kim, Jaehoon Hahm, Sumin Kim, Jinhwan Sul, Byunghak Kim, Joonseok Lee

ICCV 2025 · Video › Video generation · approach: Compositional text-to-video diffusion · 5 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 video importance denoiser with summed query to encoder (ablation)SummDiff · Mr.HiSumKendall tau: 0.071 -> 0.079Improved
replace fixed random score embeddings with learnable ones (ablation)SummDiff · Mr.HiSumKendall tau: 0.079 -> 0.086Improved
add AdaLN on time embedding (ablation)SummDiff · Mr.HiSumKendall tau: 0.086 -> 0.145Improved
add AdaLN on time and positional encodings (ablation)SummDiff · Mr.HiSumKendall tau: 0.145 -> 0.171Improved
add score quantization into codebook embeddings (ablation)SummDiff · Mr.HiSumKendall tau: 0.125 -> 0.171Improved
add classifier-free guidance (ablation)SummDiff · Mr.HiSumKendall tau: 0.171 -> 0.175Improved
add self-attention guidance (ablation)SummDiff · Mr.HiSumKendall tau: 0.175 -> 0.177Improved
increase DDIM steps from 1 to 100 (ablation)SummDiff · Mr.HiSumKendall tau: 0.170 -> 0.182Improved
increase DDIM steps from 1 to 10 (inference time) (ablation)SummDiff · Mr.HiSuminference time (ms): 11.02 -> 49.73Hurt
increase quantization bins K from 5 to 200 (ablation)SummDiff · Mr.HiSumKendall tau: 0.145 -> 0.175Improved
increase quantization bins K from 200 to 800 (ablation)SummDiff · Mr.HiSumKendall tau: 0.175 -> 0.172No clear effect
train on individual annotator scores instead of aggregated (all annotations) (ablation)SummDiff · SumMeKendall tau: 0.176 -> 0.256Improved

Later papers that compared with this method

1 later accepted paper reports SummDiff as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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