Viewpoint Equivariance for Multi-View 3D Object Detection

TL;DR

This paper introduces VEDet, a 3D object detection framework that uses multi-view geometry and viewpoint equivariance to enhance localization accuracy in multi-view camera setups, achieving state-of-the-art results.

Contribution

VEDet is the first to incorporate viewpoint equivariance through multi-view geometry and view-conditioned queries in a transformer-based 3D detection framework.

Findings

Achieves state-of-the-art performance on nuScenes benchmark.

Utilizes multi-view geometry for improved 3D localization.

Enforces multi-view consistency during training.

Abstract

3D object detection from visual sensors is a cornerstone capability of robotic systems. State-of-the-art methods focus on reasoning and decoding object bounding boxes from multi-view camera input. In this work we gain intuition from the integral role of multi-view consistency in 3D scene understanding and geometric learning. To this end, we introduce VEDet, a novel 3D object detection framework that exploits 3D multi-view geometry to improve localization through viewpoint awareness and equivariance. VEDet leverages a query-based transformer architecture and encodes the 3D scene by augmenting image features with positional encodings from their 3D perspective geometry. We design view-conditioned queries at the output level, which enables the generation of multiple virtual frames during training to learn viewpoint equivariance by enforcing multi-view consistency. The multi-view geometry injected at the input level as positional encodings and regularized at the loss level provides rich geometric cues for 3D object detection, leading to state-of-the-art performance on the nuScenes benchmark. The code and model are made available at https://github.com/TRI-ML/VEDet.