DVMN: Dense Validity Mask Network for Depth Completion
This work addresses the need for dense depth completion in applications such as autonomous navigation, representing an incremental improvement over existing methods.
The paper tackles the problem of sparse LiDAR depth maps being insufficient for tasks like autonomous navigation by proposing DVMN, a guided convolutional neural network that introduces a novel layer with spatially variant dilation and a sparsity invariant residual block, achieving state-of-the-art results on the KITTI depth completion benchmark.
LiDAR depth maps provide environmental guidance in a variety of applications. However, such depth maps are typically sparse and insufficient for complex tasks such as autonomous navigation. State of the art methods use image guided neural networks for dense depth completion. We develop a guided convolutional neural network focusing on gathering dense and valid information from sparse depth maps. To this end, we introduce a novel layer with spatially variant and content-depended dilation to include additional data from sparse input. Furthermore, we propose a sparsity invariant residual bottleneck block. We evaluate our Dense Validity Mask Network (DVMN) on the KITTI depth completion benchmark and achieve state of the art results. At the time of submission, our network is the leading method using sparsity invariant convolution.