AI Paper Atlas

Research map / Video / Video segmentation and tracking

Minimum Latency Deep Online Video Stabilization

Zhuofan Zhang, Zhen Liu, Ping Tan, Bing Zeng, Shuaicheng Liu

ICCV 2023 · Video › Video segmentation and tracking · approach: Learned optical flow estimation · 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 shape-consistency loss to motion-consistency loss (ablation)online stabilization UNet · NUS datasetDistortion value: 0.84 -> 0.90Improved
add shape-consistency loss to motion-consistency loss (ablation)online stabilization UNet · NUS datasetStability score: 0.87 -> 0.86No clear effect
add scale-preserving loss on top of MC and SC losses (ablation)online stabilization UNet · NUS datasetCropping ratio: 0.85 -> 0.90Improved
add scale-preserving loss on top of MC and SC losses (ablation)online stabilization UNet · NUS datasetDistortion value: 0.90 -> 0.95Improved
add channel attention to the path smoothing network (ablation)online stabilization UNet · NUS datasetCropping ratio: 0.90 -> 0.92Improved
add channel attention to the path smoothing network (ablation)online stabilization UNet · NUS datasetStability score: 0.86 -> 0.87No clear effect
train on MotionStab instead of DeepStab (ablation)online stabilization UNet and StabNet · NUS datasetCropping ratio and distortion valueImproved
use deep motion model plus path-smoothing network instead of Meshflowonline video stabilization · NUS dataset (Avg.)Cropping ratio: 0.7882 -> 0.9242Improved
use proposed online method instead of Meshflow on stabilityonline video stabilization · NUS dataset (Avg.)Stability score: 0.8617 -> 0.8705Improved
online method versus best offline methods on stabilityonline video stabilization · NUS dataset (Avg.)Stability score: 0.9062 -> 0.8705Hurt

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