Fine-Grained Pillar Feature Encoding Via Spatio-Temporal Virtual Grid for 3D Object Detection

TL;DR

This paper introduces FG-PFE, a novel pillar encoding method using spatio-temporal virtual grids to better capture LiDAR point distributions, significantly improving 3D object detection accuracy with minimal computational cost.

Contribution

The paper proposes a new fine-grained pillar feature encoding architecture utilizing spatio-temporal grids, enhancing point distribution capture within pillars for improved detection performance.

Findings

FG-PFE outperforms baseline models on nuScenes dataset.

Significant accuracy improvements with minimal additional computation.

Effective encoding of point distributions across vertical, temporal, and horizontal dimensions.

Abstract

Developing high-performance, real-time architectures for LiDAR-based 3D object detectors is essential for the successful commercialization of autonomous vehicles. Pillar-based methods stand out as a practical choice for onboard deployment due to their computational efficiency. However, despite their efficiency, these methods can sometimes underperform compared to alternative point encoding techniques such as Voxel-encoding or PointNet++. We argue that current pillar-based methods have not sufficiently captured the fine-grained distributions of LiDAR points within each pillar structure. Consequently, there exists considerable room for improvement in pillar feature encoding. In this paper, we introduce a novel pillar encoding architecture referred to as Fine-Grained Pillar Feature Encoding (FG-PFE). FG-PFE utilizes Spatio-Temporal Virtual (STV) grids to capture the distribution of point clouds within each pillar across vertical, temporal, and horizontal dimensions. Through STV grids, points within each pillar are individually encoded using Vertical PFE (V-PFE), Temporal PFE (T-PFE), and Horizontal PFE (H-PFE). These encoded features are then aggregated through an Attentive Pillar Aggregation method. Our experiments conducted on the nuScenes dataset demonstrate that FG-PFE achieves significant performance improvements over baseline models such as PointPillar, CenterPoint-Pillar, and PillarNet, with only a minor increase in computational overhead.