PLG-IN: Pluggable Geometric Consistency Loss with Wasserstein Distance in Monocular Depth Estimation

TL;DR

This paper introduces PLG-IN, a novel geometric consistency loss based on Wasserstein distance, to enhance monocular depth and pose estimation by penalizing inconsistencies between point clouds.

Contribution

The paper presents a pluggable Wasserstein distance-based loss function for geometric consistency, improving depth and pose estimation accuracy in monocular images.

Findings

Improved depth and pose estimation accuracy on KITTI dataset.

Effective penalization of geometric inconsistencies.

Compatibility with existing state-of-the-art methods.

Abstract

We propose a novel objective for penalizing geometric inconsistencies to improve the depth and pose estimation performance of monocular camera images. Our objective is designed using the Wasserstein distance between two point clouds, estimated from images with different camera poses. The Wasserstein distance can impose a soft and symmetric coupling between two point clouds, which suitably maintains geometric constraints and results in a differentiable objective. By adding our objective to the those of other state-of-the-art methods, we can effectively penalize geometric inconsistencies and obtain highly accurate depth and pose estimations. Our proposed method is evaluated using the KITTI dataset.