Human Instance Segmentation On Ochuman

Metrics

AP

Results

Performance results of various models on this benchmark

		Paper Title
HQNet (ViT-L)	38.8	You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception
BBox-Mask-Pose 2x	32.4	Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing the Virtuous Circle
Crowd-SAM (ViT-L)	31.4	Crowd-SAM: SAM as a Smart Annotator for Object Detection in Crowded Scenes
HQNet (ResNet-50)	31.1	You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception
BlendMask + CIS	29.8	Real-time Human-Centric Segmentation for Complex Video Scenes
Mask2Former + Occlusion C&P	28.3	Humans need not label more humans: Occlusion Copy & Paste for Occluded Human Instance Segmentation
CondInst + CIS	28.1	Real-time Human-Centric Segmentation for Complex Video Scenes
Mask2Former	27.8	Object-Centric Multi-Task Learning for Human Instances
HCQNet	27.3	Object-Centric Multi-Task Learning for Human Instances
ExPoSeg	26.8	PoSeg: Pose-Aware Refinement Network for Human Instance Segmentation
RTMDet-ins-l	26.5	Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing the Virtuous Circle
JoPoSeg	26.4	PoSeg: Pose-Aware Refinement Network for Human Instance Segmentation
BaseNet-DPS	25.5	Object-Centric Multi-Task Learning for Human Instances
Pose2Seg	23.8	Pose2Seg: Detection Free Human Instance Segmentation
PolarMask + CIS	23.4	Real-time Human-Centric Segmentation for Complex Video Scenes
ResNet-101-FPN + TTG v1	22.42	Test-time Adaptation vs. Training-time Generalization: A Case Study in Human Instance Segmentation using Keypoints Estimation
BCNet	20.6	Occlusion-Aware Instance Segmentation via BiLayer Network Architectures
CaSe	18.0	Count- and Similarity-aware R-CNN for Pedestrian Detection
Mask RCNN	16.9	Count- and Similarity-aware R-CNN for Pedestrian Detection

0 of 19 row(s) selected.

Human Instance Segmentation On Ochuman | SOTA | HyperAI