Patch Refinement -- Localized 3D Object Detection

TL;DR

Patch Refinement is a two-stage 3D object detection method that improves accuracy by focusing on localized patches, enabling higher resolution encoding and efficient training, demonstrated on the KITTI benchmark.

Contribution

The paper introduces a novel two-stage detection framework that decomposes 3D detection into BEV and local patch-based steps, enhancing resolution and efficiency.

Findings

Outperforms all previous methods on KITTI car detection benchmarks.

Uses only 50% of training data and LiDAR data, yet achieves superior results.

Enables higher voxel resolution locally through patch-based processing.

Abstract

We introduce Patch Refinement a two-stage model for accurate 3D object detection and localization from point cloud data. Patch Refinement is composed of two independently trained Voxelnet-based networks, a Region Proposal Network (RPN) and a Local Refinement Network (LRN). We decompose the detection task into a preliminary Bird's Eye View (BEV) detection step and a local 3D detection step. Based on the proposed BEV locations by the RPN, we extract small point cloud subsets ("patches"), which are then processed by the LRN, which is less limited by memory constraints due to the small area of each patch. Therefore, we can apply encoding with a higher voxel resolution locally. The independence of the LRN enables the use of additional augmentation techniques and allows for an efficient, regression focused training as it uses only a small fraction of each scene. Evaluated on the KITTI 3D object detection benchmark, our submission from January 28, 2019, outperformed all previous entries on all three difficulties of the class car, using only 50 % of the available training data and only LiDAR information.