Publications (5)
Deep Permutation Equivariant Structure from Motion
Dror Moran, Hodaya Koslowsky, Yoni Kasten +3
Existing deep methods produce highly accurate 3D reconstructions in stereo and multiview stereo settings, i.e., when cameras are both internally and externally calibrated. Neverthe…
GSVisLoc: Generalizable Visual Localization for Gaussian Splatting Scene Representations
Fadi Khatib, Dror Moran, Guy Trostianetsky +3
GSVisLoc is a visual localization technique that estimates a camera's position and orientation by matching features from a 3D Gaussian Splatting scene model to features extracted f…
Consensus Learning with Deep Sets for Essential Matrix Estimation
Dror Moran, Yuval Margalit, Guy Trostianetsky +3
Robust estimation of the essential matrix, which encodes the relative position and orientation of two cameras, is a fundamental step in structure from motion pipelines. Recent deep…
RESfM: Robust Deep Equivariant Structure from Motion
Fadi Khatib, Yoni Kasten, Dror Moran +2
Multiview Structure from Motion is a fundamental and challenging computer vision problem. A recent deep-based approach utilized matrix equivariant architectures for simultaneous re…
Multiview Neural Surface Reconstruction by Disentangling Geometry and Appearance
Lior Yariv, Yoni Kasten, Dror Moran +4
In this work we address the challenging problem of multiview 3D surface reconstruction. We introduce a neural network architecture that simultaneously learns the unknown geometry,…