atlas

Research map / Video / Video segmentation and tracking

FVC: A New Framework Towards Deep Video Compression in Feature Space

CVPR 2021 · Video › Video segmentation and tracking · approach: Learned video compression · 273 citations (OpenAlex)

Check your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 feature-space motion estimation and compensation with deformable convolution instead of pixel-space motion operations (ablation)FVC-basic · HEVC Class D, 0.4bppPSNR (dB): +1.2dBHelped
perform residual compression in feature space rather than pixel space (ablation)FVC-basic · HEVC Class D, 0.38bppPSNR (dB): +0.3dBHelped
add multi-frame feature fusion (ablation)FVC · HEVC Class D, 0.3bppPSNR (dB): +0.5dBHelped
add non-local attention to multi-frame feature fusion (ablation)FVC · HEVC Class DPSNR (dB): nearly 0.2dBHelped
use deformable feature-space compensation rather than pixel-level optical-flow compensation (ablation)FVC · HEVC Class Cpredicted-frame PSNR (dB): +1.75dB at 0.017bppHelped
use deformable compensation with offset-map coding rather than optical-flow motion coding (ablation)FVC · RaceHorses, HEVC Class CPSNR (dB): 0.27dB gain by only using 65% bppHelped

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