HyperAI超神経

Human Object Interaction Detection On V Coco

評価指標

AP(S1)
AP(S2)
Time Per Frame(ms)

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名
AP(S1)
AP(S2)
Time Per Frame(ms)
Paper TitleRepository
UPT-R10160.766.264Efficient Two-Stage Detection of Human-Object Interactions with a Novel Unary-Pairwise Transformer
MUREN68.871.0-Relational Context Learning for Human-Object Interaction Detection
DRG51.0--DRG: Dual Relation Graph for Human-Object Interaction Detection
Body Part Interactiveness63.065.1-Mining Cross-Person Cues for Body-Part Interactiveness Learning in HOI Detection
PD-Net53.34--Polysemy Deciphering Network for Robust Human-Object Interaction Detection
IDN53.360.3-HOI Analysis: Integrating and Decomposing Human-Object Interaction
QPIC (ResNet101)58.360.763QPIC: Query-Based Pairwise Human-Object Interaction Detection with Image-Wide Contextual Information
DIRV56.1-68DIRV: Dense Interaction Region Voting for End-to-End Human-Object Interaction Detection
SGCN4HOI53.157.9-A Skeleton-aware Graph Convolutional Network for Human-Object Interaction Detection
SCG54.260.9500Spatially Conditioned Graphs for Detecting Human-Object Interactions
RLIPv272.174.1-RLIPv2: Fast Scaling of Relational Language-Image Pre-training
PMN51.8--Pose-based Modular Network for Human-Object Interaction Detection
TIN (Interactiveness)49.0-513Transferable Interactiveness Knowledge for Human-Object Interaction Detection
iCAN44.7--iCAN: Instance-Centric Attention Network for Human-Object Interaction Detection
GGNet54.7--Glance and Gaze: Inferring Action-aware Points for One-Stage Human-Object Interaction Detection
TIN (TIPAMI)49.1--Transferable Interactiveness Knowledge for Human-Object Interaction Detection
OCN (ResNet50)64.266.343Detecting Human-Object Interactions with Object-Guided Cross-Modal Calibrated Semantics
HOTR55.264.4-HOTR: End-to-End Human-Object Interaction Detection with Transformers
UPT-R5059.064.543Efficient Two-Stage Detection of Human-Object Interactions with a Novel Unary-Pairwise Transformer
HOTR + CPC---Consistency Learning via Decoding Path Augmentation for Transformers in Human Object Interaction Detection
0 of 34 row(s) selected.