AI Paper Atlas

Research map / Video / Video segmentation and tracking

Event-Based Video Reconstruction Using Transformer

Wenming Weng, Yueyi Zhang, Zhiwei Xiong

ICCV 2021 · Video › Video segmentation and tracking · approach: Self-supervised video representation learning · 145 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 Token Pyramid Aggregation (vs E2VID-res12 with similar parameters) (ablation)ET-Net-4-s4 · MVSECMSE: 0.169 -> 0.118Improved
add Token Pyramid Aggregation (vs E2VID-res12 with similar parameters) (ablation)ET-Net-4-s4 · MVSECSSIM: 0.309 -> 0.355Improved
add Token Pyramid Aggregation (vs E2VID-res6 with similar parameters) (ablation)ET-Net-2-s4 · MVSECMSE: 0.165 -> 0.113Improved
use 4 aggregation scales instead of a single scale in TPA (ablation)ET-Net · IJRRMSE: 0.0991 -> 0.0552Improved
use 4 aggregation scales instead of 3 in TPA (ablation)ET-Net · IJRRMSE: 0.0584 -> 0.0552Improved
use Transformer decoders instead of encoders for cross-scale dependency in TPA (ablation)ET-Net-4-s4 · IJRRMSE: 0.0563 -> 0.0522Improved
use Transformer decoders instead of encoders for cross-scale dependency in TPA (ablation)ET-Net-4-s4 · MVSECMSE: 0.129 -> 0.118Improved
use Transformer decoders instead of encoders for cross-scale dependency in TPA (ablation)ET-Net-4-s4 · HQFMSE: 0.0399 -> 0.0403No clear effect
add skip connection between RCB and MLU (ablation)ET-Net-4-s4 · MVSECLPIPS: 0.653 -> 0.491Improved
add skip connection between RCB and MLU (ablation)E2VID-res12 · MVSECMSE: 0.179 -> 0.169Improved
use 4 Transformer encoders/decoders per scale rather than 2 (ablation)ET-Net-s4 · HQFMSE: 0.0413 -> 0.0403Improved
use 6 Transformer encoders/decoders per scale rather than 4 (ablation)ET-Net-s4 · HQFMSE: 0.0403 -> 0.0430Hurt

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