PEM360: a dataset of 360° videos with continuous physiological measurements, subjective emotional ratings and motion traces.
Quentin GuimardFlorent RobertCamille BauceAldric DucreuxLucile SassatelliHui-Yin WuMarco WincklerAuriane GrosPublished in: MMSys (2022)
Keyphrases
- human actions
- periodic motion
- human motion
- space time
- video sequences
- recognition of human actions
- motion features
- moving camera
- action recognition in videos
- input video
- motion capture data
- action recognition
- image sequences
- measurement errors
- recognizing human actions
- image measurements
- dynamic scenes
- video dataset
- key frames
- spatio temporal
- successive frames
- collaborative filtering
- motion estimation
- static images
- optical flow
- motion cues
- motion patterns
- camera motion
- motion trajectories
- motion segmentation
- motion analysis
- video frames
- netflix prize
- adjacent frames
- temporal coherence
- image motion
- multimedia event detection
- motion model
- spatial and temporal
- background subtraction
- moving objects
- video data
- user preferences
- human activities
- motion parameters
- visual data
- motion capture
- dynamic textures
- video representation
- video content
- video analysis
- global motion
- point correspondences
- traffic scenes
- video stabilization
- temporal domain
- greater degree
- visual features
- object motion