PLNet: Plane and Line Priors for Unsupervised Indoor Depth Estimation

TL;DR

PLNet introduces a novel approach for unsupervised indoor depth estimation by leveraging plane and line priors, improving accuracy in textureless regions through geometric consistency constraints.

Contribution

The paper proposes a new method that incorporates plane and line priors into unsupervised depth learning, with geometric consistency losses and evaluation metrics for flatness and straightness.

Findings

PLNet outperforms existing methods on NYU Depth V2 and ScanNet datasets.

Using plane and line priors improves depth estimation in textureless indoor scenes.

The proposed consistency losses enhance the geometric regularity of the predicted point cloud.

Abstract

Unsupervised learning of depth from indoor monocular videos is challenging as the artificial environment contains many textureless regions. Fortunately, the indoor scenes are full of specific structures, such as planes and lines, which should help guide unsupervised depth learning. This paper proposes PLNet that leverages the plane and line priors to enhance the depth estimation. We first represent the scene geometry using local planar coefficients and impose the smoothness constraint on the representation. Moreover, we enforce the planar and linear consistency by randomly selecting some sets of points that are probably coplanar or collinear to construct simple and effective consistency losses. To verify the proposed method's effectiveness, we further propose to evaluate the flatness and straightness of the predicted point cloud on the reliable planar and linear regions. The regularity of these regions indicates quality indoor reconstruction. Experiments on NYU Depth V2 and ScanNet show that PLNet outperforms existing methods. The code is available at \url{https://github.com/HalleyJiang/PLNet}.