Multi-resolution deep learning pipeline for dense large scale point clouds
This work addresses the problem of efficiently processing large-scale 3D point clouds for applications like object detection, though it appears incremental as it builds on existing multi-resolution and deep learning approaches.
The authors tackled the challenge of processing dense, large-scale 3D point clouds by introducing a multi-resolution deep learning pipeline that splits processing into sub-networks operating at different resolutions, allowing classes to benefit from either noise reduction or fine-grained details, resulting in improved handling of computational and memory costs.
Recent development of 3D sensors allows the acquisition of extremely dense 3D point clouds of large-scale scenes. The main challenge of processing such large point clouds remains in the size of the data, which induce expensive computational and memory cost. In this context, the full resolution cloud is particularly hard to process, and details it brings are rarely exploited. Although fine-grained details are important for detection of small objects, they can alter the local geometry of large structural parts and mislead deep learning networks. In this paper, we introduce a new generic deep learning pipeline to exploit the full precision of large scale point clouds, but only for objects that require details. The core idea of our approach is to split up the process into multiple sub-networks which operate on different resolutions and with each their specific classes to retrieve. Thus, the pipeline allows each class to benefit either from noise and memory cost reduction of a sub-sampling or from fine-grained details.