AI Paper Atlas

Research map / Video / Video segmentation and tracking

2.5D Visual Sound

Ruohan Gao, Kristen Grauman

CVPR 2019 · Video › Video segmentation and tracking · approach: Sign language and physiological video analysis

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 visual frames alongside mono audio rather than audio alone (ablation)Mono2Binaural · FAIR-Play, REC-STREET, YT-CLEAN, YT-MUSICSTFT distance: FAIR-Play: 0.966 -> 0.836; REC-STREET: 0.590 -> 0.565; YT-CLEAN: 1.065 -> 1.027; YT-MUSIC: 1.553 -> 1.451Improved
Use visual frames alongside mono audio rather than audio alone (ablation)Mono2Binaural · FAIR-Play, REC-STREET, YT-CLEAN, YT-MUSICENV distance: FAIR-Play: 0.141 -> 0.132; REC-STREET: 0.114 -> 0.109; YT-CLEAN: 0.131 -> 0.130Improved
Use the correct accompanying visual frames rather than flipped frames (ablation)Mono2Binaural · FAIR-Play, REC-STREET, YT-CLEAN, YT-MUSICSTFT distance: FAIR-Play: 1.145 -> 0.836; REC-STREET: 0.658 -> 0.565; YT-CLEAN: 1.095 -> 1.027; YT-MUSIC: 1.590 -> 1.451Improved
Use the correct accompanying visual frames rather than flipped frames (ablation)Mono2Binaural · YT-MUSICENV distance: 0.165 -> 0.156Improved
Use predicted binaural audio instead of mono audio for separationAudio-visual source separation · FAIR-PlaySDR: 2.57 -> 3.01Improved
Use predicted binaural audio instead of mono audio for separationAudio-visual source separation · FAIR-PlaySIR: 4.25 -> 5.03Improved
Use predicted binaural audio instead of mono audio for separationAudio-visual source separation · FAIR-PlaySAR: 10.12 -> 10.24Improved
Predict the left-right difference signal instead of directly predicting both channels (ablation)Mono2Binaural · Binaural audio generationvisual-feature useImproved

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