MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Video.
Jinlu ZhangZhigang TuJianyu YangYujin ChenJunsong YuanPublished in: CVPR (2022)
Keyphrases
- spatio temporal
- spatial and temporal
- spatial temporal
- space time
- temporal correlation
- video representation
- temporal domain
- human actions
- video data
- video encoder
- spatio temporally
- multimedia
- video sequences
- real time
- video streams
- low complexity
- video content
- video database
- encoding process
- video encoding
- temporal structure
- action detection
- real time video
- video clips
- moving objects
- video compression
- video processing
- mpeg standard
- temporal segmentation
- video transcoding
- dynamic textures
- video analysis
- video retrieval
- video surveillance
- image sequences
- video frames
- action recognition
- bit rate
- computer vision
- rate control
- three dimensional
- video copy detection
- motion estimation
- error control
- temporal information
- video shots
- spatio temporal data
- video objects