{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,7,30]],"date-time":"2025-07-30T16:31:36Z","timestamp":1753893096136,"version":"3.41.2"},"reference-count":30,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2024,10,9]],"date-time":"2024-10-09T00:00:00Z","timestamp":1728432000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Robot. AI"],"abstract":"<jats:p>Camera relocalization determines the position and orientation of a camera in a 3D space. Althouh methods based on scene coordinate regression yield highly accurate results in indoor scenes, they exhibit poor performance in outdoor scenarios due to their large scale and increased complexity. A visual localization method, Py-Net, is therefore proposed herein. Py-Net is based on voting segmentation and comprises a main encoder containing Py-layer and two branch decoders. The Py-layer comprises pyramid convolution and 1 \u00d7 1 convolution kernels for feature extraction across multiple levels, with fewer parameters to enhance the model\u2019s ability to extract scene information. Coordinate attention was added at the end of the encoder for feature correction, which improved the model robustness to interference. To prevent the feature loss caused by repetitive structures and low-texture images in the scene, deep over-parameterized convolution modules were incorporated into the seg and vote decoders. Landmark segmentation and voting maps were used to establish the relation between images and landmarks in 3D space, reducing anomalies and achieving high precision with a small number of landmarks. The experimental results show that, in multiple outdoor scenes, Py-Net achieves lower distance and angle errors compared to existing methods. Additionally, compared to VS-Net, which also uses a voting segmentation structure, Py-Net reduces the number of parameters by 31.85% and decreases the model size from 236MB to 170 MB.<\/jats:p>","DOI":"10.3389\/frobt.2024.1469588","type":"journal-article","created":{"date-parts":[[2024,10,9]],"date-time":"2024-10-09T13:01:11Z","timestamp":1728478871000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Enhanced outdoor visual localization using Py-Net voting segmentation approach"],"prefix":"10.3389","volume":"11","author":[{"given":"Jing","family":"Wang","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Cheng","family":"Guo","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shaoyi","family":"Hu","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yibo","family":"Wang","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xuhui","family":"Fan","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2024,10,9]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"2792","DOI":"10.11591\/ijece.v12i3.pp2792-2801","article-title":"Depth-DensePose: an efficient densely connected deep learning model for camera-based localization","volume":"12","author":"Abozeid","year":"2022","journal-title":"Int. J. Electr. and Comput. Eng."},{"key":"B2","first-page":"26","article-title":"CNN architecture for weakly supervised place recognition","volume-title":"Proceedings of the IEEE conference on computer vision and pattern recognition","author":"Arandjelovic","year":"2016"},{"key":"B3","first-page":"10","article-title":"Continuous metric learning relocalisation using neural nets","volume-title":"Proceedings of the European conference on computer vision","author":"Balntas","year":"2018"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52729.2023.00488","article-title":"Accelerated coordinate encoding: learning to relocalize in minutes using rgb and poses","author":"Brachmann","year":"2023","journal-title":"CVPR"},{"key":"B5","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2017.267","article-title":"Dsac-differentiable ransac for camera localization","volume-title":"Proceedings of the IEEE conference on computer vision and pattern recognition","author":"Brachmann","year":"2017"},{"key":"B6","doi-asserted-by":"publisher","first-page":"5847","DOI":"10.1109\/tpami.2021.3070754","article-title":"Visual camera Re-localization from RGB and RGB-D images using DSAC","volume":"44","author":"Brachmann","year":"","journal-title":"IEEE Trans. Pattern Analysis Mach. Intell."},{"key":"B7","doi-asserted-by":"publisher","first-page":"5847","DOI":"10.1109\/TPAMI.2021.3070754","article-title":"Visual camera re-localization from RGB and RGB-D images using DSAC","volume":"44","author":"Brachmann","year":"","journal-title":"IEEE Trans. pattern analysis Mach. Intell."},{"key":"B8","first-page":"01830","article-title":"Fast and lightweight scene regressor for camera relocalization","author":"Bui","year":"2022","journal-title":"arXiv"},{"key":"B9","first-page":"27","article-title":"Camera relocalization by exploiting multi-view constraints for scene coordinates regression","volume-title":"Proceed-ings of the IEEE\/CVF international conference on computer vision workshops","author":"Cai","year":"2019"},{"key":"B10","doi-asserted-by":"publisher","first-page":"3726","DOI":"10.1109\/tip.2022.3175432","article-title":"Do-conv: depthwise over-parameterized convolutional layer","volume":"31","author":"Cao","year":"2022","journal-title":"IEEE Trans. Image Pro-cessing"},{"key":"B11","first-page":"373","article-title":"Survey of monocular camera-based visual relocalization","volume":"43","author":"Chen","year":"2021","journal-title":"Robot"},{"key":"B12","first-page":"8540","article-title":"Robust neural routing through space partitions for camera relocalization in dynamic indoor environments","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition","author":"Dong","year":"2021"},{"key":"B13","doi-asserted-by":"publisher","first-page":"102850","DOI":"10.1016\/j.cviu.2019.102850","article-title":"Efficient multi-output scene coordinate prediction for fast and accurate camera relocali-zation from a single RGB image","volume":"190","author":"Duong","year":"2020","journal-title":"Comput. Vis. Image Underst."},{"key":"B14","first-page":"6097","article-title":"VS-net: voting with segmentation for visual localization","volume-title":"Proceedings of the IEEE\/CVF con-ference on computer vision and pattern recognition","author":"Huang","year":"2021"},{"key":"B15","first-page":"11","article-title":"Posenet: a convolutional network for real-time 6-DOF camera relocalization","volume-title":"Pro-ceedings of the IEEE international conference on computer vision","author":"Kendall","year":""},{"key":"B16","article-title":"Posenet: a convolutional network for real-time 6-dof camera relocali-zation","volume-title":"Proceedings of the IEEE international conference on computer vision","author":"Kendall","year":""},{"key":"B17","first-page":"22","article-title":"Camera relocalization by computing pairwise relative poses using convolutional neural network","volume-title":"Proceedings of the IEEE international conference on computer vision workshops","author":"Laskar","year":"2017"},{"key":"B18","article-title":"Scene coordinate regression with angle-based reprojection loss for camera relocalization","volume-title":"Proceedings of the European conference on computer vision (ECCV) workshops","author":"Li","year":"2018"},{"key":"B19","first-page":"14","article-title":"Hierarchical scene coordinate classification and regression for visual localization","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition","author":"Li","year":"2020"},{"key":"B20","doi-asserted-by":"crossref","DOI":"10.1109\/ICRA48891.2023.10161420","article-title":"Nerf-loc: visual localization ith conditional neural radiance field","volume-title":"Icra","author":"Liu","year":"2023"},{"key":"B21","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2018.00546","article-title":"Lstm pose machines","volume-title":"Proceedings of the IEEE conference on computer vision and pattern recognition","author":"Luo","year":"2018"},{"key":"B22","doi-asserted-by":"crossref","first-page":"19","DOI":"10.1109\/CVPR46437.2021.00326","article-title":"Back to the Feature: Learning Robust Camera Localization from Pixels to Pose","volume-title":"2021 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Sarlin","year":"2021"},{"key":"B23","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR46437.2021.00326","article-title":"Back to the feature: learning robust camera localization from pixels to pose","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition","author":"Sarlin","year":""},{"key":"B24","first-page":"05272","article-title":"Introduction to camera pose estimation with deep learning","author":"Shavit","year":"2019","journal-title":"arXiv"},{"key":"B25","doi-asserted-by":"crossref","DOI":"10.1109\/ICCAR.2019.8813337","article-title":"Supervised high-dimension endecoder net: 3D end to end prediction network for mark-less human pose estimation from single depth map","volume-title":"2019 5th international conference on control, automation and robotics (ICCAR)","author":"Shen","year":"2019"},{"key":"B26","first-page":"1","article-title":"Visual camera Re-localization using graph neural networks and relative pose supervision","volume-title":"Proceedings of the international conference on 3D vision","author":"Turkoglu","year":"2021"},{"key":"B27","first-page":"23","article-title":"Beyond controlled environments: 3D camera Re-localization in changing indoor scenes","volume-title":"Proceedings of the European conference on computer vision","author":"Wald","year":"2020"},{"key":"B28","doi-asserted-by":"publisher","first-page":"105294","DOI":"10.1109\/access.2024.3435851","article-title":"ST-PixLoc: a scene-agnostic network for enhanced camera localization","volume":"12","author":"Wang","year":"2024","journal-title":"IEEE Access"},{"key":"B29","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV.2019.00013","article-title":"SANet: scene agnostic network for camera localization","volume-title":"Proceedings of the IEEE\/CVF interna-tional conference on computer vision","author":"Yang","year":"2019"},{"key":"B30","first-page":"14","article-title":"KFNet: learning temporal camera relocalization using kalman filtering","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition","author":"Zhou","year":"2020"}],"container-title":["Frontiers in Robotics and AI"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2024.1469588\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,9]],"date-time":"2024-10-09T13:01:22Z","timestamp":1728478882000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2024.1469588\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,10,9]]},"references-count":30,"alternative-id":["10.3389\/frobt.2024.1469588"],"URL":"https:\/\/doi.org\/10.3389\/frobt.2024.1469588","relation":{},"ISSN":["2296-9144"],"issn-type":[{"type":"electronic","value":"2296-9144"}],"subject":[],"published":{"date-parts":[[2024,10,9]]},"article-number":"1469588"}}