Evaluation of Class Distribution and Class Combinations on Semantic Segmentation of 3D Point Clouds With PointNet

Abstract

Point clouds are generated by light imaging, detection and ranging (LIDAR) scanners or depth imaging cameras, which capture the geometry from the scanned objects with high accuracy. Unfortunately, these systems are unable to identify the semantics of the objects. Semantic 3D point clouds are an important basis for modeling the real world in digital applications. Manual semantic segmentation is a labor and cost intensive task. Automation of semantic segmentation using machine learning and deep learning (DL) approaches is therefore an interesting subject of research. In particular, point-based network architectures, such as PointNet, lead to a beneficial semantic segmentation in individual applications. For the application of DL methods, a large number of hyperparameters (HPs) have to be determined and these HPs influence the training success. In our work, the investigated HPs are the class distribution and the class combination. By means of seven combinations of classes following a hierarchical scheme and four methods to adapt the class sizes, these HPs are investigated in a detailed and structured manner. The investigated settings show an increased semantic segmentation performance, by an increase of 31% in recall for the class Erroneous points or that all classes have a recall of higher than 50%. However, based on our results the correct setting of only these HPs does not lead to a simple, universal and practical semantic segmentation procedure.

Dynamic Spatial–Spectral Feature Optimization-Based Point Cloud Classification

“…Unlike 2D detection based on image data from optical sensors, LiDAR's spatial sensing capabilities are advantageous for object recognition in 3D spaces [5]. LiDAR is a radar system that detects a target's position, velocity, and other characteristic parameters by emitting a laser beam [6]. Compared to traditional two-dimensional remote sensing data, the point cloud contains geometric information and potentially encompasses information about color or intensity [7].…”