3DFCNN: Real-Time Action Recognition using 3D Deep Neural Networks with Raw Depth Information.
Adrian Sanchez-CaballeroSergio de López DizDavid Fuentes-JiménezCristina Losada-GutiérrezMarta Marrón RomeraDavid Casillas-PerezMohammad Ibrahim SarkerPublished in: CoRR (2020)
Keyphrases
- depth information
- depth cameras
- real time
- neural network
- depth map
- stereo vision
- microsoft kinect
- depth recovery
- depth images
- stereo camera
- depth data
- rgbd images
- view synthesis
- time of flight
- image based rendering
- depth sensors
- motion parallax
- disparity map
- preprocessing
- d scene
- multi view
- video sequences
- input image
- shape from focus
- stereoscopic video
- data sets
- multi view images
- image data
- vision system
- single image