Sparse and Dense Data with CNNs: Depth Completion and Semantic Segmentation
This addresses the challenge of processing sparse vision data like lidar for autonomous driving applications, offering a robust and efficient solution.
The paper tackles the problem of handling sparse depth data, optionally combined with dense RGB, for depth completion and semantic segmentation using CNNs, achieving state-of-the-art performance on the Kitti depth completion benchmark even with densities as low as 0.8%.
Convolutional neural networks are designed for dense data, but vision data is often sparse (stereo depth, point clouds, pen stroke, etc.). We present a method to handle sparse depth data with optional dense RGB, and accomplish depth completion and semantic segmentation changing only the last layer. Our proposal efficiently learns sparse features without the need of an additional validity mask. We show how to ensure network robustness to varying input sparsities. Our method even works with densities as low as 0.8% (8 layer lidar), and outperforms all published state-of-the-art on the Kitti depth completion benchmark.