Occ$^2$Net: Robust Image Matching Based on 3D Occupancy Estimation for Occluded Regions
This addresses a critical issue for visual applications like SLAM and image retrieval by improving matching accuracy in occluded areas, though it is an incremental advancement in handling occlusions.
The paper tackles the problem of image matching in occluded regions by proposing Occ^2Net, which models occlusion relations using 3D occupancy estimation, resulting in superior performance over state-of-the-art methods in occlusion scenarios.
Image matching is a fundamental and critical task in various visual applications, such as Simultaneous Localization and Mapping (SLAM) and image retrieval, which require accurate pose estimation. However, most existing methods ignore the occlusion relations between objects caused by camera motion and scene structure. In this paper, we propose Occ$^2$Net, a novel image matching method that models occlusion relations using 3D occupancy and infers matching points in occluded regions. Thanks to the inductive bias encoded in the Occupancy Estimation (OE) module, it greatly simplifies bootstrapping of a multi-view consistent 3D representation that can then integrate information from multiple views. Together with an Occlusion-Aware (OA) module, it incorporates attention layers and rotation alignment to enable matching between occluded and visible points. We evaluate our method on both real-world and simulated datasets and demonstrate its superior performance over state-of-the-art methods on several metrics, especially in occlusion scenarios.