CNN-based synthesis of realistic high-resolution LiDAR data
This work addresses the need for high-quality LiDAR data synthesis in domains like autonomous driving, though it appears incremental as it builds on existing CNN techniques with novel loss functions.
The paper tackles the problem of synthesizing realistic high-resolution LiDAR point cloud data using a CNN-based approach with specialized loss functions, achieving significant quantitative and qualitative improvements in geometry and semantics over traditional non-CNN methods.
This paper presents a novel CNN-based approach for synthesizing high-resolution LiDAR point cloud data. Our approach generates semantically and perceptually realistic results with guidance from specialized loss-functions. First, we utilize a modified per-point loss that addresses missing LiDAR point measurements. Second, we align the quality of our generated output with real-world sensor data by applying a perceptual loss. In large-scale experiments on real-world datasets, we evaluate both the geometric accuracy and semantic segmentation performance using our generated data vs. ground truth. In a mean opinion score testing we further assess the perceptual quality of our generated point clouds. Our results demonstrate a significant quantitative and qualitative improvement in both geometry and semantics over traditional non CNN-based up-sampling methods.