AI Paper Atlas

Research map / Video / Video segmentation and tracking

Decompose the Sounds and Pixels, Recompose the Events

Varshanth R. Rao, Md Ibrahim Khalil, Haoda Li, Peng Dai, Juwei Lu

AAAI 2022 · Video › Video segmentation and tracking · approach: Sign language and physiological video analysis · 6 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
use EDRNet instead of PSPAVE localization · O-AVE WSELaccuracy (%): 72.80 -> 73.86Improved
use EDRNet instead of PSPAVE localization · O-AVE SELaccuracy (%): 76.84 -> 77.93Improved
add B2ILC (ablation)EDRNet · O-AVE SELaccuracy (%): 76.10 -> 76.58Improved
add B2ILC (ablation)EDRNet · O-AVE WSELaccuracy (%): 72.98 -> 73.86Improved
add LSS loss (ablation)EDRNet + B2ILC · O-AVE SELaccuracy (%): 76.58 -> 77.10Improved
add SMB video fusion (ablation)EDRNet + B2ILC + LSS loss · O-AVE SELaccuracy (%): 77.10 -> 77.93Improved
remove positional encoding (ablation)EDRNet · O-AVE WSELaccuracy (%): 73.86% -> 72.88% (↓0.98%)Hurt
remove positional encoding (ablation)EDRNet · O-AVE SELaccuracy (%): 77.93% -> 77.09% (↓0.84%)Hurt
add the AV decomposition branch to audio-only (ablation)EDRNet audio branch · O-AVE SELaccuracy (%): 65.84 -> 76.11Improved
add the AV decomposition branch to visual-only (ablation)EDRNet visual branch · O-AVE SELaccuracy (%): 66.93 -> 75.26Improved
enable audio and visual branches with late fusion, without the AV branch (ablation)EDRNet · O-AVE SELaccuracy (%): 75.88Improved
enable audio, visual, and AV decomposition branches (ablation)EDRNet · O-AVE SELaccuracy (%): 75.88 -> 77.93Improved

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