AI Paper Atlas

Research map / Video / Video segmentation and tracking

Many-to-Many Splatting for Efficient Video Frame Interpolation

Ping Hu, Simon Niklaus, Stan Sclaroff, Kate Saenko

CVPR 2022 · Video › Video segmentation and tracking · approach: Video frame interpolation · 73 citations (OpenAlex)

Check 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 M2M splatting with PWC-Net flowVideo frame interpolation · X-TEST(4K), ×8 interpolationPSNR: 30.16 -> 30.81Improved
Use M2M splatting with PWC-Net flowVideo frame interpolation · X-TEST(2K), ×8 interpolationPSNR: 30.65 -> 32.07Improved
Add a Motion Refinement Network to the initial flow (ablation)M2M-PWC · Vimeo90KPSNR: 33.97 -> 34.94Improved
Add a Motion Refinement Network to the initial flow (ablation)M2M-DIS · Vimeo90KPSNR: 31.93 -> 34.32Improved
Add Joint Flow Encoding (ablation)M2M-PWC with MRN · Vimeo90KPSNR: 34.94 -> 35.09Improved
Add Low-rank Feature Modulation (ablation)M2M-PWC with MRN · Vimeo90KPSNR: 34.94 -> 35.07Improved
Add both Joint Flow Encoding and Low-rank Feature Modulation (ablation)M2M-PWC with MRN · Vimeo90KPSNR: 34.94 -> 35.15Improved
Add both Joint Flow Encoding and Low-rank Feature Modulation (ablation)M2M-DIS with MRN · Vimeo90KPSNR: 34.32 -> 34.78Improved
Add the learned reliability score to pixel fusion (ablation)M2M-PWC with JFE and LFM · Vimeo90KPSNR: 35.15 -> 35.24Improved
Add the learned reliability score to pixel fusion (ablation)M2M-DIS with JFE and LFM · Vimeo90KPSNR: 34.78 -> 34.93Improved
Increase sub-motion vectors from 1 to 4 per pixel (ablation)M2M-PWC · Vimeo90KPSNR: 35.24 -> 35.40Improved
Increase sub-motion vectors from 1 to 4 per pixel (ablation)M2M-DIS · Vimeo90KPSNR: 34.93 -> 35.06Improved

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