ConQueR: Query Contrast Voxel-DETR for 3D Object Detection

TL;DR

ConQueR introduces a contrastive query mechanism in a sparse 3D detector to significantly reduce false positives and improve detection accuracy, achieving state-of-the-art results on the Waymo dataset.

Contribution

The paper proposes a novel Query Contrast mechanism for sparse 3D detection, explicitly discriminating queries to enhance accuracy and reduce false positives.

Findings

Reduces false positives by up to 60%.

Achieves 71.6 mAPH/L2 on Waymo dataset, surpassing previous methods.

Outperforms PV-RCNN++ by over 2.0 mAPH/L2.

Abstract

Although DETR-based 3D detectors can simplify the detection pipeline and achieve direct sparse predictions, their performance still lags behind dense detectors with post-processing for 3D object detection from point clouds. DETRs usually adopt a larger number of queries than GTs (e.g., 300 queries v.s. 40 objects in Waymo) in a scene, which inevitably incur many false positives during inference. In this paper, we propose a simple yet effective sparse 3D detector, named Query Contrast Voxel-DETR (ConQueR), to eliminate the challenging false positives, and achieve more accurate and sparser predictions. We observe that most false positives are highly overlapping in local regions, caused by the lack of explicit supervision to discriminate locally similar queries. We thus propose a Query Contrast mechanism to explicitly enhance queries towards their best-matched GTs over all unmatched query predictions. This is achieved by the construction of positive and negative GT-query pairs for each GT, and a contrastive loss to enhance positive GT-query pairs against negative ones based on feature similarities. ConQueR closes the gap of sparse and dense 3D detectors, and reduces up to ~60% false positives. Our single-frame ConQueR achieves new state-of-the-art (sota) 71.6 mAPH/L2 on the challenging Waymo Open Dataset validation set, outperforming previous sota methods (e.g., PV-RCNN++) by over 2.0 mAPH/L2.