Depth completion starts from a sparse set of known depth values and estimates the unknown depths for the remaining image pixels. Most methods model this as depth interpolation and erroneously interpolate depth pixels into the empty space between spatially distinct objects, resulting in depth-smearing across occlusion boundaries. Here we propose a multi-hypothesis depth representation that explicitly models both foreground and background depths in the difficult occlusion-boundary regions. Our method can be thought of as performing twin-surface extrapolation, rather than interpolation, in these regions. Next our method fuses these extrapolated surfaces into a single depth image leveraging the image data. Key to our method is the use of an asymmetric loss function that operates on a novel twin-surface representation. This enables us to train a network to simultaneously do surface extrapolation and surface fusion. We characterize our loss function and compare with other common losses. Finally, we validate our method on three different datasets; KITTI, an outdoor real-world dataset, NYU2, indoor real-world depth dataset and Virtual KITTI, a photo-realistic synthetic dataset with dense groundtruth, and demonstrate improvement over the state of the art.
Figure 1: Our depth completion algorithm can input LiDAR data and image (a), and extrapolate the estimates of foreground depth d1 (b) and background depth d2 (c), along with a weight σ (e). Fusing all three leads to the completed depth (d). The foreground-background depth difference (f) d2-d1 is small except at depth discontinuities.
KITTI Comparison
Figure 2: Comparison of our method with SoTA methods with whole and zoom in views (a) showing Color Images (b) DC (c) MultiStack (d) NLSPN and our method (e). Four different regions of the image from two different instants are selected to show depth quality from diverse areas.