Domain-invariant Stereo Matching Networks
This addresses the domain generalization issue in stereo matching for computer vision applications, representing a strong specific gain rather than an incremental improvement.
The paper tackles the problem of stereo matching networks struggling to generalize to unseen environments due to domain differences, and the result is a domain-invariant network (DSMNet) that significantly outperforms state-of-the-art models on real test sets, even surpassing some fine-tuned models.
State-of-the-art stereo matching networks have difficulties in generalizing to new unseen environments due to significant domain differences, such as color, illumination, contrast, and texture. In this paper, we aim at designing a domain-invariant stereo matching network (DSMNet) that generalizes well to unseen scenes. To achieve this goal, we propose i) a novel "domain normalization" approach that regularizes the distribution of learned representations to allow them to be invariant to domain differences, and ii) a trainable non-local graph-based filter for extracting robust structural and geometric representations that can further enhance domain-invariant generalizations. When trained on synthetic data and generalized to real test sets, our model performs significantly better than all state-of-the-art models. It even outperforms some deep learning models (e.g. MC-CNN) fine-tuned with test-domain data.