ACSFNet: Adaptive Channel-Spatial Fusion for Efficient 2D Human Pose Estimation
ACSFNet introduces adaptive channel-spatial fusion to improve 2D human pose estimation efficiency via ACSFNet-18/30, ACT, SDG, and SimCC implementations with benchmarking utilities.
Published Jul 17, 2026Paper ↗
67%
OverallHighly rated
?
OverallHighly ratedVote to see the score
Readers
–
Only vote on papers you've read. Sign in to vote.
AI panel3/20reviewers recommend it
lenient 1/5
medium 2/10
strict 0/5
AI panel?Vote to see what the 20 AI reviewers said
Abstract
Reference implementation corresponding to the revised ACSFNet manuscript. This release includes ACSFNet-18/30, ACT, SDG, SimCC, training and evaluation scripts, ablation configurations, and efficiency benchmarking utilities.