A Simple Method to Boost Human Pose Estimation Accuracy by Correcting the Joint Regressor for the Human3.6m Dataset.
Eric HedlinHelge RhodinKwang Moo YiPublished in: CRV (2022)
Keyphrases
- estimation accuracy
- human pose
- human actions
- pose estimation
- action recognition
- human object interactions
- body parts
- human body
- prediction accuracy
- human motion
- human detection
- estimation error
- benchmark datasets
- human activities
- computer vision
- spatio temporal
- latent variables
- three dimensional
- sample size
- pedestrian detection
- visual features
- depth images
- image classification
- feature extraction