Depth Based Semantic Scene Completion With Position Importance Aware Loss.
Jie LiYu LiuXia YuanChunxia ZhaoRoland SiegwartIan ReidCesar CadenaPublished in: IEEE Robotics Autom. Lett. (2020)
Keyphrases
- defocused images
- d scene
- depth map
- depth estimation
- stereoscopic images
- real scenes
- scene geometry
- three dimensional
- imaging sensors
- semantic context
- relative position
- input image
- positional information
- natural language
- depth information
- depth cues
- precise localization
- video sequences
- multiple images
- semantic information
- stereo camera
- semantic web
- single image
- dynamic scenes
- depth images
- moving objects
- scene understanding
- spatial location
- semantic similarity
- low level features
- scene classification
- multi view
- single shot
- relative depth