Cascaded Pyramid Network for Multi-Person Pose Estimation

The topic of multi-person pose estimation has been largely improved recently, especially with the development of convolutional neural network. However, there still exist a lot of challenging cases, such as occluded keypoints, invisible keypoints and complex background, which cannot be well addressed... (read more)

PDF Abstract CVPR 2018 PDF CVPR 2018 Abstract

Datasets


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Multi-Person Pose Estimation COCO CPN+ AP 0.730 # 2
Keypoint Detection COCO CPN+ Test AP 73.0 # 7
Keypoint Detection COCO test-challenge CPN+ AR 78.7 # 3
ARM 74.3 # 3
AP 72.1 # 3
AP50 90.5 # 3
AP75 78.9 # 3
APL 84.7 # 3
AR50 94.7 # 3
AR75 84.8 # 3
ARL 78.1 # 3
Keypoint Detection COCO test-dev CPN APL 77.2 # 6
APM 68.7 # 7
AP50 91.4 # 7
AP75 80.0 # 7
AR 78.5 # 6
AR50 95.1 # 3
AR75 85.3 # 4
ARL 84.3 # 4
ARM 74.2 # 4
Keypoint Detection COCO test-dev CPN+ APL 78.1 # 5
APM 69.5 # 6
AP50 91.7 # 6
AP75 80.9 # 6
AR 79.0 # 5
AR50 95.1 # 3
AR75 85.9 # 3
ARL 84.6 # 3
ARM 74.8 # 3
Pose Estimation COCO test-dev CPN AP 72.1 # 10
AP50 91.4 # 7
AP75 80.0 # 9
APL 77.2 # 9
AR 78.5 # 7
Pose Estimation COCO test-dev CPN+ [6, 9] AP 73.0 # 8
AP50 91.7 # 6
AP75 80.9 # 8
APL 78.1 # 8
AR 79.0 # 6

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet