single mention · not yet filed
Block-wise depth masking for depth-completion evaluation
Block-wise Depth Masking is an evaluation protocol for depth completion that creates incomplete depth maps by randomly masking out spatial regions of varying sizes from the ground-truth depth.1 It simulates the depth dropout seen in consumer depth cameras, and the remaining depth values can be corrupted with noise to form difficulty levels from easy to extreme.2
How this definition was written
Summarised by glm-5.3-flash from 1 source; every sentence is checked against the quote it cites.
We generate incomplete depth maps by randomly masking out spatial regions of varying sizes (blocks) from the ground-truth depth
— Masked Depth Modeling for Spatial Perceptionsimulating depth dropout commonly observed in consumer depth cameras. To further model low-quality depth measurements, we corrupt the remaining depth values with additive Gaussian noise
— Masked Depth Modeling for Spatial Perception
Also called Block-wise Depth Masking.