SemanticPOSS 3D Semantic Segmentation Dataset

SemanticPOSS is a dataset for studying 3D semantic segmentation. This dataset provides new methods for 3D semantic segmentation, the main task in autonomous driving research. The dataset contains 2,988 LiDAR scans of various complexities and a large number of dynamic instances. The data was collected at Peking University and uses the same data format as SemanticKITTI.