Densely Constrained Depth Estimator for Monocular 3D Object Detection

TL;DR

This paper introduces a novel monocular 3D object detection method that leverages dense projection constraints from edges in any direction, significantly improving depth estimation accuracy and achieving state-of-the-art results.

Contribution

It proposes a densly constrained depth estimator utilizing edges of any direction and a graph matching weighting module to merge depth candidates, advancing monocular 3D detection.

Findings

Achieves state-of-the-art performance on KITTI and WOD benchmarks.

Utilizes dense projection constraints for improved depth candidate generation.

Employs a graph matching module for effective depth candidate merging.

Abstract

Estimating accurate 3D locations of objects from monocular images is a challenging problem because of lacking depth. Previous work shows that utilizing the object's keypoint projection constraints to estimate multiple depth candidates boosts the detection performance. However, the existing methods can only utilize vertical edges as projection constraints for depth estimation. So these methods only use a small number of projection constraints and produce insufficient depth candidates, leading to inaccurate depth estimation. In this paper, we propose a method that utilizes dense projection constraints from edges of any direction. In this way, we employ much more projection constraints and produce considerable depth candidates. Besides, we present a graph matching weighting module to merge the depth candidates. The proposed method DCD (Densely Constrained Detector) achieves state-of-the-art performance on the KITTI and WOD benchmarks. Code is released at https://github.com/BraveGroup/DCD.