AI Paper Atlas

Research map / Video / Action recognition

Spatio-Temporal Proximity-Aware Dual-Path Model for Panoramic Activity Recognition

Sumin Lee, Yooseung Wang, Sangmin Woo, Changick Kim

ECCV 2024 · Video › Action recognition · approach: Skeleton-based action recognition · 3 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 spatial (height and width) attention on top of temporal attention (ablation)SPDP-Net individual self-attention · JRDB-PAROverall F1: 40.3 -> 46.5Improved
add any spatio-temporal attention versus none (ablation)SPDP-Net · JRDB-PAROverall F1: 37.4 -> 46.5Improved
use K-means with predicted number of groups instead of spectral clustering (ablation)SPDP-Net social group detection · JRDB-PARIoU@0.5: 49.1 -> 56.4Improved
add auxiliary loss L_aux on individual attention (ablation)SPDP-Net · JRDB-PARSocial group detection IoU@0.5: 51.8 -> 53.3Improved
add relation loss L_R alone (ablation)SPDP-Net · JRDB-PARSocial group detection IoU@0.5: 51.8 -> 51.4No clear effect
combine L_aux and L_R (ablation)SPDP-Net · JRDB-PARSocial group detection IoU@0.5: 53.3 -> 56.4Improved
increase relation loss weight to 3 (ablation)SPDP-Net · JRDB-PAROverall F1: 46.5 -> 44.9Hurt
increase individual or auxiliary loss weight (ablation)SPDP-Net · JRDB-PARSocial activity F1: F_p 34.2 -> 29.8 (3:1:1)Mixed

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