AdaDepth: Unsupervised Content Congruent Adaptation for Depth Estimation
This addresses the challenge of generalizing depth estimation models from synthetic to natural scenes without costly ground truth, though it is incremental as it builds on existing adversarial methods.
The paper tackles the problem of domain shift in monocular depth estimation by proposing AdaDepth, an unsupervised domain adaptation strategy that uses adversarial learning and content consistency, achieving competitive results in unsupervised settings and state-of-the-art in semi-supervised settings.
Supervised deep learning methods have shown promising results for the task of monocular depth estimation; but acquiring ground truth is costly, and prone to noise as well as inaccuracies. While synthetic datasets have been used to circumvent above problems, the resultant models do not generalize well to natural scenes due to the inherent domain shift. Recent adversarial approaches for domain adaption have performed well in mitigating the differences between the source and target domains. But these methods are mostly limited to a classification setup and do not scale well for fully-convolutional architectures. In this work, we propose AdaDepth - an unsupervised domain adaptation strategy for the pixel-wise regression task of monocular depth estimation. The proposed approach is devoid of above limitations through a) adversarial learning and b) explicit imposition of content consistency on the adapted target representation. Our unsupervised approach performs competitively with other established approaches on depth estimation tasks and achieves state-of-the-art results in a semi-supervised setting.