360-MLC: Multi-view Layout Consistency for Self-training and Hyper-parameter Tuning

TL;DR

360-MLC introduces a self-training approach leveraging multi-view layout consistency in 360-images to adapt monocular room-layout models to new domains without ground truth annotations.

Contribution

It proposes a novel multi-view consistency-based self-training method for monocular layout estimation that also enables hyper-parameter tuning without labeled data.

Findings

Achieves competitive performance compared to state-of-the-art methods.

Effectively uses uncertainty and entropy for pseudo-label weighting and model evaluation.

Demonstrates successful domain adaptation on a new multi-view dataset.

Abstract

We present 360-MLC, a self-training method based on multi-view layout consistency for finetuning monocular room-layout models using unlabeled 360-images only. This can be valuable in practical scenarios where a pre-trained model needs to be adapted to a new data domain without using any ground truth annotations. Our simple yet effective assumption is that multiple layout estimations in the same scene must define a consistent geometry regardless of their camera positions. Based on this idea, we leverage a pre-trained model to project estimated layout boundaries from several camera views into the 3D world coordinate. Then, we re-project them back to the spherical coordinate and build a probability function, from which we sample the pseudo-labels for self-training. To handle unconfident pseudo-labels, we evaluate the variance in the re-projected boundaries as an uncertainty value to weight each pseudo-label in our loss function during training. In addition, since ground truth annotations are not available during training nor in testing, we leverage the entropy information in multiple layout estimations as a quantitative metric to measure the geometry consistency of the scene, allowing us to evaluate any layout estimator for hyper-parameter tuning, including model selection without ground truth annotations. Experimental results show that our solution achieves favorable performance against state-of-the-art methods when self-training from three publicly available source datasets to a unique, newly labeled dataset consisting of multi-view of the same scenes.