It's All Around You: Range-Guided Cylindrical Network for 3D Object Detection
This work addresses the problem of efficiently processing 3D LiDAR data for autonomous driving perception systems by proposing a more suitable coordinate system and adaptive convolutions, offering an incremental improvement for researchers and developers in this domain.
This paper introduces a novel 3D object detection network that processes LiDAR data in a cylindrical coordinate system, aligning with the sensor's scanning pattern. It also incorporates range-guided convolutions that adapt the receptive field based on distance and object scale, achieving results comparable to state-of-the-art architectures on the nuScenes challenge.
Modern perception systems in the field of autonomous driving rely on 3D data analysis. LiDAR sensors are frequently used to acquire such data due to their increased resilience to different lighting conditions. Although rotating LiDAR scanners produce ring-shaped patterns in space, most networks analyze their data using an orthogonal voxel sampling strategy. This work presents a novel approach for analyzing 3D data produced by 360-degree depth scanners, utilizing a more suitable coordinate system, which is aligned with the scanning pattern. Furthermore, we introduce a novel notion of range-guided convolutions, adapting the receptive field by distance from the ego vehicle and the object's scale. Our network demonstrates powerful results on the nuScenes challenge, comparable to current state-of-the-art architectures. The backbone architecture introduced in this work can be easily integrated onto other pipelines as well.