Denoising 3D Human Poses from Low-Resolution Video using Variational Autoencoder.
Chihiro NakatsukaSatoshi KomoritaPublished in: IROS (2021)
Keyphrases
- low resolution
- denoising
- depth images
- human pose
- low frame rate
- high resolution
- lower resolution
- super resolution
- higher resolution
- human actions
- image denoising
- pose estimation
- action recognition
- human body
- high quality
- image processing
- video sequences
- video data
- human detection
- image patches
- multimedia
- body parts
- video surveillance
- depth map
- face images
- video frames
- natural images
- space time
- three dimensional
- image segmentation
- optical flow
- computer vision
- spatio temporal
- depth information
- similarity measure
- graphical models