论文标题
Semanticposs：一个具有大量动态实例的点云数据集
SemanticPOSS: A Point Cloud Dataset with Large Quantity of Dynamic Instances
论文作者
论文摘要
3D语义细分是自动驾驶系统的关键任务之一。最近，针对3D语义细分任务的深度学习模型进行了广泛的研究，但通常需要大量的培训数据。但是，目前用于3D语义分割的数据集缺乏点注释，分散场景和动态对象。 在本文中，我们提出了Semanticposs数据集，其中包含2988种具有大量动态实例的各种且复杂的激光扫描。数据是在北京大学收集的，并使用与Semantickitti相同的数据格式。此外，我们在Semanticposs数据集上评估了几种典型的3D语义分割模型。实验结果表明，Semanticposs可以帮助提高动态物体作为人的预测准确性，在某种程度上。 Semanticposs将在\ url {www.poss.pku.edu.cn}上发布。
3D semantic segmentation is one of the key tasks for autonomous driving system. Recently, deep learning models for 3D semantic segmentation task have been widely researched, but they usually require large amounts of training data. However, the present datasets for 3D semantic segmentation are lack of point-wise annotation, diversiform scenes and dynamic objects. In this paper, we propose the SemanticPOSS dataset, which contains 2988 various and complicated LiDAR scans with large quantity of dynamic instances. The data is collected in Peking University and uses the same data format as SemanticKITTI. In addition, we evaluate several typical 3D semantic segmentation models on our SemanticPOSS dataset. Experimental results show that SemanticPOSS can help to improve the prediction accuracy of dynamic objects as people, car in some degree. SemanticPOSS will be published at \url{www.poss.pku.edu.cn}.
