Attention Unet++ for lightweight depth estimation from sparse depth samples and a single RGB image.
Tao ZhaoShuguo PanWang GaoChao ShengYingchun SunJiansheng WeiPublished in: Vis. Comput. (2022)
Keyphrases
- depth estimation
- lightweight
- depth map
- stereo vision
- stereo matching
- depth information
- depth cues
- input image
- scene understanding
- image data
- real scenes
- dynamic scenes
- stereo pair
- motion parallax
- single image
- depth data
- high resolution
- feature matching
- super resolution
- multi view
- image retrieval
- disparity map
- image segmentation
- depth from defocus
- d scene
- feature points
- object detection
- three dimensional
- stereo images
- geometric constraints
- object boundaries
- post processing
- color images
- high quality
- computer vision
- depth estimates