{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,18]],"date-time":"2025-12-18T14:22:23Z","timestamp":1766067743922,"version":"3.37.3"},"reference-count":89,"publisher":"Wiley","license":[{"start":{"date-parts":[[2023,4,22]],"date-time":"2023-04-22T00:00:00Z","timestamp":1682121600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Basic Research Program of China","doi-asserted-by":"publisher","award":["2020YFB1313304"],"award-info":[{"award-number":["2020YFB1313304"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Journal of Robotics"],"published-print":{"date-parts":[[2023,4,22]]},"abstract":"<jats:p>Visual place recognition (VPR) is considered among the most challenging problems due to the extreme variations in appearance and viewpoint. Essentially, appearance-based VPR can be considered as an image retrieval task, thus the key is to accurately and efficiently describe the images. Recently, global descriptor methods have attracted substantial attention from the VPR community, which has contributed to numerous important outcomes. Despite the growing number of global descriptors presented, little attention has been paid to the comparison and evaluation of these methods and so it remains difficult for researchers to disentangle the factors that led to better performance. This study provided comprehensive insight into global descriptors from a practical application perspective. We present a systematic evaluation that integrates 15 commonly used global descriptors, 6 benchmark datasets, and 5 evaluation metrics, and subsequently extended this evaluation to discuss the key factors impacting the matching performance and computational efficiency. We also report practical suggestions for constructing promising CNN descriptors, based on the experimental conclusions. Our analysis reveals both advantages and limitations of three different types of global descriptors, including handcrafted features-based ones, off-the-shelf CNN-based ones, and customized CNN-based ones. Finally, we evaluate the practicality of reported global descriptors to mediate the trade-offs between matching performance and computational efficiency.<\/jats:p>","DOI":"10.1155\/2023\/9150357","type":"journal-article","created":{"date-parts":[[2023,4,22]],"date-time":"2023-04-22T23:20:14Z","timestamp":1682205614000},"page":"1-17","source":"Crossref","is-referenced-by-count":4,"title":["Evaluation of Global Descriptor Methods for Appearance-Based Visual Place Recognition"],"prefix":"10.1155","volume":"2023","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-9266-2340","authenticated-orcid":true,"given":"Kangyu","family":"Li","sequence":"first","affiliation":[{"name":"Machinery Technology Development Co., Ltd., Beijing 100037, China"},{"name":"China Academy of Machinery Science and Technology, Beijing 100044, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yuhan","family":"Ma","sequence":"additional","affiliation":[{"name":"Machinery Technology Development Co., Ltd., Beijing 100037, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xifeng","family":"Wang","sequence":"additional","affiliation":[{"name":"China Academy of Machinery Science and Technology, Beijing 100044, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Lijuan","family":"Ji","sequence":"additional","affiliation":[{"name":"China Productivity Center for Machinery, Beijing 101407, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Niuniu","family":"Geng","sequence":"additional","affiliation":[{"name":"Machinery Technology Development Co., Ltd., Beijing 100037, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","reference":[{"issue":"2","key":"1","doi-asserted-by":"crossref","first-page":"99","DOI":"10.1109\/MRA.2006.1638022","article-title":"Simultaneous localization and mapping: part i","volume":"13","author":"H. Durrant-Whyte","year":"2006","journal-title":"IEEE Robotics and Automation Magazine"},{"key":"2","doi-asserted-by":"publisher","DOI":"10.15607\/RSS.2015.XI.022"},{"key":"3","doi-asserted-by":"publisher","DOI":"10.1177\/0278364908090961"},{"key":"4","doi-asserted-by":"publisher","DOI":"10.1109\/tro.2012.2197158"},{"key":"5","doi-asserted-by":"publisher","DOI":"10.1109\/tro.2015.2496823"},{"key":"6","doi-asserted-by":"publisher","DOI":"10.1109\/ROBOT.2000.844029"},{"key":"7","doi-asserted-by":"publisher","DOI":"10.1109\/mgrs.2021.3081763"},{"key":"8","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2005.6"},{"key":"9","doi-asserted-by":"publisher","DOI":"10.1016\/S0079-6123(06)55002-2"},{"first-page":"2161","article-title":"Scalable recognition with a vocabulary tree","author":"D. Nister","key":"10"},{"key":"11","doi-asserted-by":"publisher","DOI":"10.1023\/b:visi.0000029664.99615.94"},{"first-page":"2564","article-title":"Orb: an efficient alternative to sift or surf","author":"E. Rublee","key":"12"},{"key":"13","doi-asserted-by":"publisher","DOI":"10.1109\/IROS33365.2015"},{"key":"14","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR33180.2016"},{"key":"15","doi-asserted-by":"publisher","DOI":"10.1016\/j.cag.2017.07.019"},{"key":"16","doi-asserted-by":"publisher","DOI":"10.1007\/s10514-015-9516-2"},{"key":"17","doi-asserted-by":"publisher","DOI":"10.1109\/IROS37595.2017"},{"article-title":"Levelling the playing field: a comprehensive comparison of visual place recognition approaches under changing conditions","year":"2019","author":"M. Zaffar","key":"18"},{"key":"19","doi-asserted-by":"publisher","DOI":"10.1007\/s10462-012-9365-8"},{"key":"20","doi-asserted-by":"publisher","DOI":"10.1109\/icra14741.2009"},{"key":"21","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2020.107760"},{"key":"22","doi-asserted-by":"publisher","DOI":"10.1109\/tro.2008.2004514"},{"key":"23","doi-asserted-by":"publisher","DOI":"10.1109\/tro.2015.2463671"},{"key":"24","doi-asserted-by":"publisher","DOI":"10.1109\/tro.2017.2705103"},{"first-page":"5979","article-title":"Assigning visual words to places for loop closure detection","author":"K. A. Tsintotas","key":"25"},{"first-page":"3212","article-title":"Feature co-occurrence maps: appearance-based localisation throughout the day","author":"E. Johns","key":"26"},{"first-page":"5571","article-title":"Condition-invariant, top-down visual place recognition","author":"M. Milford","key":"27"},{"key":"28","doi-asserted-by":"publisher","DOI":"10.15607\/RSS.2014.X.023"},{"key":"29","doi-asserted-by":"publisher","DOI":"10.1109\/lra.2020.2969917"},{"first-page":"2196","article-title":"Experiments in place recognition using gist panoramas","author":"A. C. Murillo","key":"30"},{"first-page":"4042","article-title":"Visual loop closing using gist descriptors in manhattan world","author":"G. Singh","key":"31"},{"key":"32","doi-asserted-by":"publisher","DOI":"10.1109\/tro.2012.2192013"},{"first-page":"5441","article-title":"Ibuild: incremental bag of binary words for appearance based loop closure detection","author":"S. Khan","key":"33"},{"first-page":"1","article-title":"City-scale location recognition","author":"G. Schindler","key":"34"},{"volume-title":"k-means++: The Advantages of Careful Seeding","year":"2006","author":"D. Arthur","key":"35"},{"key":"36","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR12085.2007"},{"first-page":"3304","article-title":"Aggregating local descriptors into a compact image representation","author":"H. J\u00e9gou","key":"37"},{"article-title":"Convolutional neural network-based place recognition","year":"2014","author":"Z. Chen","key":"38"},{"key":"39","doi-asserted-by":"publisher","DOI":"10.1109\/ICIA36264.2015"},{"key":"40","doi-asserted-by":"publisher","DOI":"10.1007\/s10846-018-0917-2"},{"key":"41","doi-asserted-by":"publisher","DOI":"10.1109\/access.2020.2982228"},{"key":"42","article-title":"Imagenet classification with deep convolutional neural networks","volume":"25","author":"A. Krizhevsky","year":"2012","journal-title":"Advances in Neural Information Processing Systems"},{"key":"43","doi-asserted-by":"publisher","DOI":"10.1109\/tpami.2017.2723009"},{"article-title":"Very deep convolutional networks for large-scale image recognition","year":"2014","author":"K. Simonyan","key":"44"},{"key":"45","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR33180.2016"},{"key":"46","doi-asserted-by":"crossref","DOI":"10.15607\/RSS.2018.XIV.032","article-title":"Lightweight unsupervised deep loop closure","author":"N. Merrill","year":"2018"},{"key":"47","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2017.12.069"},{"first-page":"3223","article-title":"Deep learning features at scale for visual place recognition","author":"Z. Chen","key":"48"},{"key":"49","doi-asserted-by":"publisher","DOI":"10.1016\/j.patrec.2017.04.017"},{"key":"50","doi-asserted-by":"publisher","DOI":"10.1109\/5.726791"},{"first-page":"1","article-title":"Going deeper with convolutions","author":"C. Szegedy","key":"51"},{"first-page":"1251","article-title":"Xception: deep learning with depthwise separable convolutions","author":"F. Chollet","key":"52"},{"first-page":"4700","article-title":"Densely connected convolutional networks","author":"G. Huang","key":"53"},{"article-title":"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and< 0.5 MB model size","year":"2016","author":"F. N. Iandola","key":"54"},{"article-title":"Mobilenets: efficient convolutional neural networks for mobile vision applications","year":"2017","author":"A. G. Howard","key":"55"},{"first-page":"1314","article-title":"Searching for mobilenetv3","author":"A. Howard","key":"56"},{"first-page":"6848","article-title":"Shufflenet: an extremely efficient convolutional neural network for mobile devices","author":"X. Zhang","key":"57"},{"first-page":"6105","article-title":"Efficientnet: rethinking model scaling for convolutional neural networks","author":"M. Tan","key":"58"},{"key":"59","doi-asserted-by":"publisher","DOI":"10.1109\/tro.2021.3075644"},{"first-page":"3268","article-title":"Filter early, match late: improving network-based visual place recognition","author":"S. Hausler","key":"60"},{"key":"61","doi-asserted-by":"publisher","DOI":"10.1109\/tro.2019.2956352"},{"first-page":"5979","article-title":"Localizing discriminative visual landmarks for place recognition","author":"Z. Xin","key":"62"},{"first-page":"116","article-title":"Shufflenet v2: practical guidelines for efficient cnn architecture design","author":"N. Ma","key":"63"},{"first-page":"595","article-title":"Fast and robust visual loop closure detection with convolutional neural network","author":"M. Zhu","key":"64"},{"key":"65","doi-asserted-by":"publisher","DOI":"10.3390\/s21134499"},{"key":"66","doi-asserted-by":"publisher","DOI":"10.24251\/HICSS.2020.069"},{"key":"67","doi-asserted-by":"publisher","DOI":"10.1117\/1.jei.31.6.061816"},{"first-page":"378","article-title":"Fast and incremental loop closure detection using proximity graphs","author":"S. An","key":"68"},{"key":"69","doi-asserted-by":"publisher","DOI":"10.21817\/indjcse\/2022\/v13i5\/221305101"},{"key":"70","doi-asserted-by":"publisher","DOI":"10.1109\/lra.2021.3094228"},{"article-title":"Ghost-dil-NetVLAD: a lightweight neural network for visual place recognition","year":"2021","author":"Q. Gong","key":"71"},{"key":"72","doi-asserted-by":"publisher","DOI":"10.1109\/jas.2022.105635"},{"first-page":"1218","article-title":"Seqlpd: sequence matching enhanced loop-closure detection based on large-scale point cloud description for self-driving vehicles","author":"Z. Liu","key":"73"},{"first-page":"14141","article-title":"Patch-netvlad: multi-scale fusion of locally-global descriptors for place recognition","author":"S. Hausler","key":"74"},{"key":"75","doi-asserted-by":"publisher","DOI":"10.1109\/tnnls.2019.2908982"},{"first-page":"456","article-title":"Leveraging deep visual descriptors for hierarchical efficient localization","author":"P.-E. Sarlin","key":"76"},{"first-page":"12716","article-title":"From coarse to fine: robust hierarchical localization at large scale","author":"P.-E. Sarlin","key":"77"},{"first-page":"224","article-title":"Superpoint: self-supervised interest point detection and description","author":"D. DeTone","key":"78"},{"first-page":"3787","article-title":"Robust loop closure detection based on bag of SuperPoints and graph verification","author":"H. Yue","key":"79"},{"key":"80","doi-asserted-by":"publisher","DOI":"10.1126\/science.1127647"},{"key":"81","doi-asserted-by":"publisher","DOI":"10.1007\/s12555-019-0891-x"},{"article-title":"Auto-encoding variational bayes","year":"2013","author":"D. P. Kingma","key":"82"},{"first-page":"4554","article-title":"CALC2. 0: combining appearance, semantic and geometric information for robust and efficient visual loop closure","author":"N. Merrill","key":"83"},{"key":"84","doi-asserted-by":"publisher","DOI":"10.1109\/lra.2018.2849609"},{"first-page":"5702","article-title":"Fast-seqslam: a fast appearance based place recognition algorithm","author":"S. M. Siam","key":"85"},{"author":"N. S\u00fcnderhauf","key":"86","article-title":"Are we there yet? challenging seqslam on a 3000 km journey across all four seasons"},{"volume-title":"Day and Night with Lateral Pose Change Datasets","year":"2014","author":"A. Glover","key":"87"},{"first-page":"9532","article-title":"A cross-season correspondence dataset for robust semantic segmentation","author":"M. Larsson","key":"88"},{"key":"89","doi-asserted-by":"publisher","DOI":"10.1109\/ICRA15123.2012"}],"container-title":["Journal of Robotics"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/downloads.hindawi.com\/journals\/jr\/2023\/9150357.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"http:\/\/downloads.hindawi.com\/journals\/jr\/2023\/9150357.xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"http:\/\/downloads.hindawi.com\/journals\/jr\/2023\/9150357.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,4,22]],"date-time":"2023-04-22T23:20:20Z","timestamp":1682205620000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.hindawi.com\/journals\/jr\/2023\/9150357\/"}},"subtitle":[],"editor":[{"given":"Xianfeng","family":"Yuan","sequence":"additional","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2023,4,22]]},"references-count":89,"alternative-id":["9150357","9150357"],"URL":"https:\/\/doi.org\/10.1155\/2023\/9150357","relation":{},"ISSN":["1687-9619","1687-9600"],"issn-type":[{"type":"electronic","value":"1687-9619"},{"type":"print","value":"1687-9600"}],"subject":[],"published":{"date-parts":[[2023,4,22]]}}}