AI Paper Atlas

Research map / Video / Video generation

Keep-or-Drop? Adaptive Tokenizer for Compact Video Representation

Yeonkyeong Lee, Hyunsung Go, Jongmin Kim, Sewoong Lim, Donghoon Lee

ECCV 2026 · Video › Video generation · approach: Efficient video diffusion models · 0 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
use finer decoder patches than encoder (asymmetric coarse-to-fine decoding) (ablation)KATok · Panda-70M val reconstructionPSNR: 29.61 -> 31.26Improved
raise sparsity loss weight from 0.01 to 0.02 (ablation)KATok stage 1 · Panda-70M valPSNR: 28.78 -> 22.46Hurt
raise sparsity loss weight from 0.01 to 0.02 (ablation)KATok stage 1 · Panda-70M valrFVD: 50.38 -> 462.27Hurt
lower sparsity weight from 0.01 to 0.005 (ablation)KATok stage 1 · Panda-70M valPSNR: 28.78 -> 28.67No clear effect
use KATok adaptive tokenizer instead of OmniTokenizer-VAE under same SiT-XLSiT-XL diffusion · UCF-101 generationgFVD: 100.00 -> 61.53Improved
use KATok instead of ElasticTok-KL under same SiT-XLSiT-XL diffusion · UCF-101 generationgFVD: 712.56 -> 61.53Improved
flip a fraction of ground-truth mask cells before content generation (1-2%) (ablation)KATok cascaded generation · Kinetics-600 generationgFVD: 69.58 -> 72.63 (1%), 81.26 (2%)Hurt
flip 10% of mask cells (ablation)KATok cascaded generation · Kinetics-600 generationgFVD: 69.58 -> 264.77Hurt
reduce image patch size from 32^2 to 16^2 (more initial tokens) (ablation)KATok on ImageNet · ImageNet val reconstructionPSNR: 25.38 -> 27.94Improved
use predicted masks from mask prior instead of random maskscascaded KATok · Kinetics-600 mask FVDmask FVD: 15,405.39 -> 91.80Improved

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