Centre de documentation
scientifique

DeepSim-Nets: Deep Similarity Networks for stereo image matching / Mohamed Ali Chebbi (2023)  

Public Titre : DeepSim-Nets: Deep Similarity Networks for stereo image matching Type de document : Article/Communication Auteurs : Mohamed Ali Chebbi, Auteur ; Ewelina Rupnik , Auteur ; Marc Pierrot-Deseilligny , Auteur ; Paul Lopes, Auteur Editeur : Computer vision foundation CVF Année de publication : 2023 Conférence : CVPR 2023, IEEE Conference on Computer Vision and Pattern Recognition 18/06/2023 22/06/2023 Vancouver Colombie britannique - Canada OA Proceedings Importance : pp 2096 - 2104 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] appariement d'images [Termes IGN] chaîne de traitement [Termes IGN] géométrie de l'image [Termes IGN] géométrie épipolaire [Termes IGN] réseau neuronal profond Index. décimale : 35.20 Traitement d'image Résumé : (auteur) We present three multi-scale similarity learning architectures, or DeepSim networks. These models learn pixel-level matching with a contrastive loss and are agnostic to the geometry of the considered scene. We establish a middle ground between hybrid and end-to-end approaches by learning to densely allocate all corresponding pixels of an epipolar pair at once. Our features are learnt on large image tiles to be expressive and capture the scene's wider context. We also demonstrate that curated sample mining can enhance the overall robustness of the predicted similarities and improve the performance on radiometrically homogeneous areas. We run experiments on aerial and satellite datasets. Our DeepSim-Nets outperform the baseline hybrid approaches and generalize better to unseen scene geometries than end-to-end methods. Our flexible architecture can be readily adopted in standard multi-resolution image matching pipelines. The code is available at https://github.com/DaliCHEBBI/DeepSimNets. Numéro de notice : C2023-007 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Thématique : IMAGERIE/INFORMATIQUE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : sans En ligne : https://openaccess.thecvf.com/content/CVPR2023W/EarthVision/html/Chebbi_DeepSim- [...] Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=103281

Pointless global bundle adjustment with relative motions Hessians / Ewelina Rupnik (2023)  

Public Titre : Pointless global bundle adjustment with relative motions Hessians Type de document : Article/Communication Auteurs : Ewelina Rupnik , Auteur ; Marc Pierrot-Deseilligny , Auteur Editeur : Computer vision foundation CVF Année de publication : 2023 Conférence : CVPR 2023, IEEE Conference on Computer Vision and Pattern Recognition 18/06/2023 22/06/2023 Vancouver Colombie britannique - Canada OA Proceedings Importance : pp 6517 - 6525 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Photogrammétrie numérique [Termes IGN] compensation par faisceaux [Termes IGN] estimation de pose [Termes IGN] matrice Index. décimale : 33.30 Photogrammétrie numérique Résumé : (auteur) Bundle adjustment (BA) is the standard way to optimise camera poses and to produce sparse representations of a scene. However, as the number of camera poses and features grows, refinement through bundle adjustment becomes inefficient. Inspired by global motion averaging methods, we propose a new bundle adjustment objective which does not rely on image features' reprojection errors yet maintains precision on par with classical BA. Our method averages over relative motions while implicitly incorporating the contribution of the structure in the adjustment. To that end, we weight the objective function by local hessian matrices-a by-product of local bundle adjustments performed on relative motions (eg, pairs or triplets) during the pose initialisation step. Such hessians are extremely rich as they encapsulate both the features' random errors and the geometric configuration between the cameras. These pieces of information propagated to the global frame help to guide the final optimisation in a more rigorous way. We argue that this approach is an upgraded version of the motion averaging approach and demonstrate its effectiveness on both photogrammetric datasets and computer vision benchmarks. Numéro de notice : C2023-008 Affiliation des auteurs : UGE-LASTIG (2020- ) Autre URL associée : vers OA paper Thématique : IMAGERIE/INFORMATIQUE/MATHEMATIQUE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : sans En ligne : https://openaccess.thecvf.com/content/CVPR2023W/PCV/papers/Rupnik_Pointless_Glob [...] Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=103282