IC-FPS: Instance-Centroid Faster Point Sampling Framework for 3D Point-based Object Detection

3D object detection is one of the most important tasks in autonomous driving and robotics. Our research focuses on tackling low efficiency issue of point-based methods, and we propose a novel Instance-Centroid Faster Point Sampling (IC-FPS) framework. We design a Neighboring Feature Diffusion Module (NFDM) to extract local features for the purpose of efficiently distinguishing the foreground from the background. Considering Farthest Point Sampling (FPS) strategy for downsampling is computationally intensive, we propose the Centroid-Instance Sampling Strategy (CISS). CISS samples center point in large-scale point cloud by rapidly sampling the centroid and instance points of the foreground block. The proposed IC-FPS framework can be inserted into every point-based model and effectively replace the first Set Abstraction (SA) layer. Extensive experiments on several public benchmarks demonstrate the superior performance of our proposed IC-FPS. On the Waymo dataset, IC-FPS significantly improves performance of the benchmark model and increases inference speed by 3. 8 times. And real-time detection of point-based methods is realized for the first time, which is meaningful for industrial applications.

Authors

• Haotian Hu
• Fanyi Wang
• Yaonong Wang
• Laifeng Hu
• Zhiwang Zhang

Keywords

• Point cloud compression
• Solid modeling
• Three-dimensional displays
• Computational modeling
• Object detection
• Benchmark testing
• Feature extraction
• Real-time systems
• Intelligent robots
• Autonomous vehicles
• Sampling Framework
• 3D Detection
• 3D Object Detection
• Central Point
• Point Cloud
• Neighborhood Characteristics
• Inference Speed
• Centroid Point
• Point-based Methods
• Learning Rate
• Layer-by-layer
• Bounding Box
• Original Information
• Feature Aggregation
• Nearest Point
• Distance Constraints
• 3D Convolution
• Centroid Position
• KITTI Dataset
• Feature Block
• Original Point Cloud
• Raw Point Cloud
• Voxel-based Methods
• Point Cloud Features