Focal Length and Object Pose Estimation via Render and Compare

TL;DR

This paper presents FocalPose, a neural render-and-compare approach that jointly estimates camera focal length and object 6D pose from a single RGB image, improving accuracy over existing methods.

Contribution

It introduces a focal length update rule for joint pose and focal length estimation and explores effective loss functions for this task.

Findings

Lower error in focal length and pose estimates compared to state-of-the-art.

Effective combination of regression and reprojection loss improves accuracy.

Validated on three challenging benchmark datasets.

Abstract

We introduce FocalPose, a neural render-and-compare method for jointly estimating the camera-object 6D pose and camera focal length given a single RGB input image depicting a known object. The contributions of this work are twofold. First, we derive a focal length update rule that extends an existing state-of-the-art render-and-compare 6D pose estimator to address the joint estimation task. Second, we investigate several different loss functions for jointly estimating the object pose and focal length. We find that a combination of direct focal length regression with a reprojection loss disentangling the contribution of translation, rotation, and focal length leads to improved results. We show results on three challenging benchmark datasets that depict known 3D models in uncontrolled settings. We demonstrate that our focal length and 6D pose estimates have lower error than the existing state-of-the-art methods.