PFENet: Towards precise feature extraction from sparse point cloud for 3D object detection

作     者：Li, Yaochen Li, Qiao Gao, Cong Gao, Shengjing Wu, Hao Liu, Rui 

作者机构：Xi An Jiao Tong Univ Sch Software Engn Xian 710049 Peoples R China 

出 版 物：《NEURAL NETWORKS》 (Neural Netw.)

年 卷 期：2025年第185卷

页      面：107144页

核心收录：

学科分类：1002[医学-临床医学] 1001[医学-基础医学(可授医学、理学学位)] 0812[工学-计算机科学与技术（可授工学、理学学位）] 10[医学] 

基　　金：National Natural Science Foundation of China, NSFC, (62473307) National Natural Science Foundation of China, NSFC Jiangsu Provincial Key Research and Development Program, (2023-YBGY-033) Jiangsu Provincial Key Research and Development Program 

主　　题：3D object detection Sparse point cloud Autonomous driving 

摘      要：Accurate 3D point cloud object detection is crucially important for autonomous driving vehicles. The sparsity of point clouds in 3D scenes, especially for smaller targets like pedestrians and bicycles that contain fewer points, makes detection particularly challenging. To solve this problem, we propose a single-stage voxel-based 3D object detection method, namely PFENet. Firstly, we design a robust voxel feature encoding network that incorporates a stacked triple attention mechanism to enhance the extraction of key features and suppress noise. Moreover, a 3D sparse convolution layer dynamically adjusts feature processing based on output location importance, improving small object recognition. Additionally, the attentional feature fusion module in the region proposal network merges low-level spatial features with high-level semantic features, and broadens the receptive field through atrous spatial pyramid pooling to capture multi-scale features. Finally, we develop multiple detection heads for more refined feature extraction and object classification, as well as more accurate bounding box regression. Experimental results on the KITTI dataset demonstrate the effectiveness of the proposed method.