PSMNet: Position-aware Stereo Merging Network for Room Layout Estimation

TL;DR

This paper introduces PSMNet, a deep learning approach for accurate room layout estimation from stereo panoramas, leveraging a novel transformer and projection layer to improve performance in complex spaces.

Contribution

The paper presents PSMNet, a new end-to-end network with a transformer and projection layer that effectively estimates room layouts from stereo panoramas, handling noisy poses.

Findings

Outperforms state-of-the-art layout estimators

Effective in large and complex room spaces

Handles noisy pose data

Abstract

In this paper, we propose a new deep learning-based method for estimating room layout given a pair of 360 panoramas. Our system, called Position-aware Stereo Merging Network or PSMNet, is an end-to-end joint layout-pose estimator. PSMNet consists of a Stereo Pano Pose (SP2) transformer and a novel Cross-Perspective Projection (CP2) layer. The stereo-view SP2 transformer is used to implicitly infer correspondences between views, and can handle noisy poses. The pose-aware CP2 layer is designed to render features from the adjacent view to the anchor (reference) view, in order to perform view fusion and estimate the visible layout. Our experiments and analysis validate our method, which significantly outperforms the state-of-the-art layout estimators, especially for large and complex room spaces.