Sparse PointPillars: Maintaining and Exploiting Input Sparsity to Improve Runtime on Embedded Systems

Bird's Eye View (BEV) is a popular representation for processing 3D point clouds, and by its nature is fundamentally sparse. Motivated by the computational limitations of mobile robot platforms, we create a fast, high-performance BEV 3D object detector that maintains and exploits this input sparsity to decrease runtimes over non-sparse baselines and avoids the tradeoff between pseudoimage area and runtime. We present results on KITTI, a canonical 3D detection dataset, and Matterport-Chair, a novel Matterport3D-derived chair detection dataset from scenes in real furnished homes. We evaluate runtime characteristics using a desktop GPU, an embedded ML accelerator, and a robot CPU, demonstrating that our method results in significant detection speedups (2 × or more) for embedded systems with only a modest decrease in detection quality. Our work represents a new approach for practitioners to optimize models for embedded systems by maintaining and exploiting input sparsity throughout their entire pipeline to reduce runtime and resource usage while preserving detection performance. All models, weights, experimental configurations, and datasets used are publicly available 1 1 https://vedder.io/sparse_point_pillars.

Authors

• Kyle Vedder
• Eric Eaton

Keywords

• Point cloud compression
• Runtime
• Three-dimensional displays
• Embedded systems
• Pipelines
• Graphics processing units
• Detectors
• Input Sparsity
• Object Detection
• Point Cloud
• Resource Usage
• Bird’s Eye
• 3D Detection
• 3D Object Detection
• Machine Learning Models
• Intersection Over Union
• Bounding Box
• Sparse Matrix
• Autonomous Vehicles
• Average Precision
• Medium Density
• Nonzero Entries
• Real Task
• Quantification Model
• 3D Convolution
• Prior Art
• KITTI Dataset
• Convolution Type
• Pipelining
• 3D Bounding Box
• Service Robots
• Input Density
• AMD Ryzen
• Sparse Format
• Detection Task
• Processing Pipeline
• Identification Code