SeqSLAM: Visual route-based navigation for sunny summer days and stormy winter nights
The DOI matches its Crossref record; 20 of the 22 references it cites were found in a registry or an index.
Abstract
Learning and then recognizing a route, whether travelled during the day or at night, in clear or inclement weather, and in summer or winter is a challenging task for state of the art algorithms in computer vision and robotics. In this paper, we present a new approach to visual navigation under changing conditions dubbed SeqSLAM. Instead of calculating the single location most likely given a current image, our approach calculates the best candidate matching location within every local navigation sequence. Localization is then achieved by recognizing coherent sequences of these “local best matches”. This approach removes the need for global matching performance by the vision front-end - instead it must only pick the best match within any short sequence of images. The approach is applicable over environment changes that render traditional feature-based techniques ineffective. Using two car-mounted camera datasets we demonstrate the effectiveness of the algorithm and compare it to one of the most successful feature-based SLAM algorithms, FAB-MAP. The perceptual change in the datasets is extreme; repeated traverses through environments during the day and then in the middle of the night, at times separated by months or years and in opposite seasons, and in clear weather and extremely heavy rain. While the feature-based method fails, the sequence-based algorithm is able to match trajectory segments at 100% precision with recall rates of up to 60%.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“This has led to develop new methods that explicitly account for such variations by matching sequences [173], gathering different visual appearances into a unified representation [48], or using spatial as well as appearance information [106].”
From Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016) · cites this paper as [173]
Citations
Cites
Object recognition from local scale-invariant features
Lowe, D. G. (1999) · Proceedings of the Seventh IEEE International Conference on Computer Vision · doi:10.1109/iccv.1999.790410 · Cited by 16,472 · cited 3×
“Many place recognition techniques rely on feature-finding algorithms such as SIFT [9] and SURF [10] which, despite their impressive rotation and scale invariant properties, are inherently unsuitable when dealing with extreme perceptual M.J. Milford and G.F. Wyeth are with the School of Electrical Engineering and Compu…”
MethodSURF: Speeded Up Robust Features
Bay, H., Tuytelaars, T., & Van Gool, L. (2006) · Lecture Notes in Computer Science · doi:10.1007/11744023_32 · Cited by 14,955
MonoSLAM: Real-time single camera SLAM
Davison, A. J., Reid, I. D., Molton, N. D., & Stasse, O. (2007) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2007.1049 · Cited by 3,959 · cited 1×
“isual recognition of places and place sequences in the field of robotic navigation has advanced rapidly in recent years, with several techniques [1-5] now being competitive with laser range-based techniques.”
BackgroundA solution to the simultaneous localization and map building (SLAM) problem
Dissanayake, M. W. M. G., Newman, P., Clark, S., Durrant-Whyte, H. F., & Csorba, M. (2001) · IEEE Transactions on Robotics and Automation · doi:10.1109/70.938381 · Cited by 2,661
Matching Local Self-Similarities across Images and Videos
Shechtman, E., & Irani, M. (2007) · 2007 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2007.383198 · Cited by 1,076 · cited 1×
“One promising step [15] towards reducing the reliance on consistent feature detection uses the self-similarity of objects within a frame or video sequence to robustly recognize objects under varying conditions.”
BackgroundIncremental mapping of large cyclic environments
Gutmann, J.-S., & Konolige, K. (2003) · Proceedings 1999 IEEE International Symposium on Computational Intelligence in Robotics and Automation. CIRA'99 (Cat. No.99EX375) · doi:10.1109/cira.1999.810068 · Cited by 568 · cited 1×
“There have been attempts to increase the robustness of visual techniques to environment change by performing sensor fusion with lasers at the level of landmarks [13] or dense depth maps [14].”
BackgroundFrameSLAM: From Bundle Adjustment to Real-Time Visual Mapping
Konolige, K., & Agrawal, M. (2008) · IEEE Transactions on Robotics · doi:10.1109/tro.2008.2004832 · Cited by 532
RatSLAM: a hippocampal model for simultaneous localization and mapping
Milford, M. J., Wyeth, G. F., & Prasser, D. (2004) · IEEE International Conference on Robotics and Automation, 2004. Proceedings. ICRA '04. 2004 · doi:10.1109/robot.2004.1307183 · Cited by 399 · cited 6×
“The approach taken in this work is partly inspired by the author's previous work in vision-based mapping in RatSLAM [16] [17] [18].”
MethodInfluentialSelf-citationMapping a Suburb With a Single Camera Using a Biologically Inspired SLAM System
Milford, M. J., & Wyeth, G. F. (2008) · IEEE Transactions on Robotics · doi:10.1109/tro.2008.2004520 · Cited by 378 · cited 7×
“The approach taken in this work is partly inspired by the author's previous work in vision-based mapping in RatSLAM [16] [17] [18].”
BackgroundMethodInfluentialSelf-citationDynamic Maps for Long-Term Operation of Mobile Service Robots
Biber, P., & Duckett, T. (2005) · Robotics: Science and Systems I · doi:10.15607/rss.2005.i.003 · Cited by 334 · cited 1×
“The problem of visual place recognition across time has received far less attention, with relatively few attempts to address the problem [6, 7], including high dynamic range approaches [8].”
Background
Cited by
Items 601–700 of 987
LiPo-LCD: Combining Lines and Points for Appearance-based Loop Closure DetectionPreprint
Company-Corcoles, J. P., Garcia‐Fidalgo, E., & Ortiz, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2009.09897 · Cited by 6
“Other approaches opt for using holistic image representations [4, 26, 36], which can be faster to compute but less tolerant to visual changes, while, lately, solutions based on Convolutional Neural Networks (CNNs) [2, 3, 37] have shown to exhibit enhanced robustness and general performance,…”
Deep learning based monocular visual-inertial odometry
Esfahani, M. A. (2020) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/152280 · Cited by 6
Unsupervised mapping and semantic user localisation from first-person monocular video
Süveges, T., & McKenna, S. J. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110923 · Cited by 5
SeqLGPR: A Sequential Subsurface Feature-Based Framework for Vehicle Place Recognition
Zhang, P., Zhi, S., Chen, X., Bi, B., Xu, Z., Chen, Y., Shen, L. C., Jin, T., & Huang, X. (2025) · IEEE Transactions on Radar Systems · doi:10.1109/trs.2025.3599141 · Cited by 5
“SeqSLAM [27], which computes frame-to-frame similarity, has become a foundational method for sequence-based localization, inspiring deep learning extensions such as SeqMatchNet [17].”
Ensemble-Based Event Camera Place Recognition Under Varying Illumination
Joseph, T., Fischer, T., & Milford, M. J. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3641119 · Cited by 5
“4) Extensive evaluation on two challenging long-term driving datasets, BrisbaneEventVPR [11] and NSAVP [23] under varied lighting conditions.”
Self-citationVision-based Posture Estimation and Null Space Control for Redundant Manipulator
Oda, N., & Fujinaga, N. (2012) · IEEJ Journal of Industry Applications · doi:10.1541/ieejjia.2.48 · Cited by 5
A Sequence-Based Neuronal Model for Mobile Robot Localization
Neubert, P., Ahmad, S., & Protzel, P. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-00111-7_11 · Cited by 5
“[17] propose to search for linear segments of high similarity in this matrix (this also involves a local contrast normalization).”
BackgroundInfluentialZhengl, Z., Chen, Y., Hua, B., & Yeung, S.-K. (2023) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros55552.2023.10342102 · Cited by 5
“As for the target domain data, we adopted the ACDC dataset [22], Dark Zurich dataset [23], the nighttime images in the BDD100K dataset [24], Nighttime Driving [25] and Alderley dataset [26].”
MethodSPOT: Point Cloud Based Stereo Visual Place Recognition for Similar and Opposing Viewpoints
Carmichael, S., Agrawal, R., Vasudevan, R., & Skinner, K. A. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10610293 · Cited by 5
“For instance, SeqSLAM searches for sequences of query-reference pairs with the knowledge that revisiting an area in the same direction will produce queries in the same order as their matching references [25].”
BackgroundMethodInfluentialPRIOR-SLAM: Enabling Visual SLAM for Loop Closure Under Large Viewpoint Variations
Wang, Y., Bai, W., Zhang, Z., Wang, H., Sun, H. X., & Cao, Q. (2024) · IEEE Transactions on Robotics · doi:10.1109/tro.2024.3508193 · Cited by 6
Towards Vision-Based Pose- and Condition-Invariant Place Recognition along Routes
Pepperell, E., Corke, P., & Milford, M. J. (2014) · QUT ePrints (Queensland University of Technology) · Cited by 5
Self-citationVisual Localization in the Presence of Appearance Changes Using the Partial Order Kernel
Abdollahyan, M., Cascianelli, S., Bellocchio, E., Costante, G., Ciarfuglia, T. A., Bianconi, F., Smeraldi, F., & Fravolini, M. L. (2018) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco.2018.8553252 · Cited by 5
“We refer to these methods as NetFlow and CNN+SeqSLAM, respectively.”
MethodInfluentialVisual place recognition based on multilevel descriptors for the visually impaired people
Fang, Y., Wang, K., Cheng, R., Yang, K., & Bai, J. Y. (2019) · Target and Background Signatures · doi:10.1117/12.2532524 · Cited by 5
Visual Localization under Appearance Change: A Filtering Approach
Doan, A.-D., Latif, Y., Chin, T.-J., Liu, Y., Chng, S.-F., Do, T.-T., & Reid, I. D. (2019) · International Conference on Digital Image Computing: Techniques and Applications (DICTA) · doi:10.1109/dicta47822.2019.8945810 · Cited by 5
“Although there is a section in [2] on “multi-image queries”, it is a relatively small part of the benchmark, with only two methods tested; SeqSLAM [5] and an extension of Active Search [18] that uses triplets of images.”
BackgroundInfluentialFast and Consistent Matching for Landmark-Based Place Recognition
Liu, Q., & Duan, F. (2020) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-019-01108-9 · Cited by 5
“Other methods, such as SeqSLAM [15], used a global description of images and exploit sequential information to perform place recognition; however, global descriptors are sensitive to viewpoint changes.”
MethodDay and Night Place Recognition Based on Low-quality Night-time Images
Liu, L., Cappelle, C., & Ruichek, Y. (2020) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc45102.2020.9294480 · Cited by 5
“Facing this big challenge, tremendous progress has been made in visual place recognition [1], [2], [4], [7], [10]–[12], [14] in recent years.”
BackgroundTian, X., Yi, P., Zhang, F., Lei, J., & Hong, Y. (2022) · Sensors · doi:10.3390/s22228604 · Cited by 7
Locality-constrained continuous place recognition for SLAM in extreme conditions
Yadav, R., Pani, V., Mishra, A., Tiwari, N., & Kala, R. (2023) · Applied Intelligence · doi:10.1007/s10489-022-04415-1 · Cited by 5
“The second set of competing methods were the place recognition algorithm from which FAB-MAP [4], SeqSLAM [22], DeepSeqSLAM [2] and Patch-NetVLAD [9] were taken.”
MethodPapapetros, I. T., Kansizoglou, I., Bampis, L., & Gasteratos, A. C. (2023) · Machines · doi:10.3390/machines11050558 · Cited by 6
Cabrera, J. J., Céspedes, O. J., Cebollada, S., Reinoso, Ó., & Payá, L. (2024) · Evolving Systems · doi:10.1007/s12530-024-09604-6 · Cited by 5
“Bai et al. [21] propose the SeqCNNSLAM method, which consists in using the pre-trained AlexNet [2] to extract features and feed the SeqSLAM algorithm [22].”
MethodArshad, S., & Kim, G.-W. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3452158 · Cited by 5
“Finally, we compare the performance of SLGD-Loop with several state-of-the-art methods [4], [6], [12], [21], [27], [38] over the range of challenging datasets.”
InfluentialPatterns of approximated localised moments for visual loop closure detection
Erhan, C., Sarıyanidi, E., Şencan, O., & Temeltaş, H. (2017) · IET Computer Vision · doi:10.1049/iet-cvi.2016.0237 · Cited by 5
Dynamic Environments Localization via Dimensions Reduction of Deep Learning Features
Zhang, H., Wang, X., Du, X., Liu, M., & Chen, Q. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-68345-4_22 · Cited by 5
“We perform the evaluations on public available Norland datasets used in SeqSLAM [20].”
BackgroundMethodInfluentialCity-scale continuous visual localization
López-Antequera, M., Petkov, N., & González-Jiménez, J. (2017) · European Conference on Mobile Robots · doi:10.1109/ecmr.2017.8098692 · Cited by 6
“In this line, SeqSLAM and its extensions [9], [10], [11] pose the problem as a sequence to sequence matching procedure, obtaining good results even with drastic appearance changes due to changing seasons.”
BackgroundCamera Localization in Outdoor Garden Environments Using Artificial Landmarks
Strisciuglio, N., Vallina, M. L., Petkov, N., & Muñoz‐Salinas, R. (2018) · IEEE International Work Conference on Bioinspired Intelligence (IWOBI) · doi:10.1109/iwobi.2018.8464139 · Cited by 5
Place Recognition in Semi-Dense Maps
Ye, Y., Cieslewski, T., Loquercio, A., & Scaramuzza, D. (2017) · British Machine Vision Conference (BMVC) · doi:10.5244/c.31.74 · Cited by 5
Autonomous Urban Localization and Navigation with Limited Information
Chipka, J. B., & Campbell, M. (2018) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2018.8500693 · Cited by 5
“Similarly, SeqSLAM is another SLAM technique that aims to allow for visual navigation despite changing environmental conditions [8].”
MethodFast Monocular Vision-based Railway Localization for Situations with Varying Speeds
Li, C.-T., & Siu, W.-C. (2018) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.23919/apsipa.2018.8659660 · Cited by 5
“With local and global descriptors to describe sequences of images, their method has outperformed SeqSLAM [4], a benchmark of visual localization systems, on the Nordland dataset [25].”
BackgroundMethodInfluentialLandmarkBoost: Efficient visualContext Classifiers for Robust Localization
Dymczyk, M., Gilitschenski, I., Nieto, J. I., Lynen, S., Zeisl, B., & Siegwart, R. (2018) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594100 · Cited by 5
Semi-Supervised Deep Vision-Based Localization Using Temporal Correlation Between Consecutive Frames
Li, C.-T., Siu, W.-C., & Lun, D. P.-K. (2019) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2019.8803131 · Cited by 5
“Alderley Dataset This dataset is described in [3] which focuses on extreme changes in weather and lighting conditions.”
MethodInfluentialEarly Bird: Loop Closures from Opposing Viewpoints for Perceptually-aliased Indoor Environments
Tourani, S., Desai, D., Parihar, U. S., Garg, S., Sarvadevabhatla, R. K., Milford, M. J., & Krishna, K. M. (2021) · International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP) · doi:10.5220/0010230804090416 · Cited by 5
Self-citationWaheed, M., Milford, M. J., McDonald-Maier, K. D., & Ehsan, S. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981722 · Cited by 12
Self-citationFaster robot perception using Salient Depth Partitioning
Chan, D. M., Taylor, A. M., & Riek, L. D. (2017) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2017.8206275 · Cited by 10
“low-power, low-computational algorithms for mobile robots, such as UV sky segmentation [46], RatSLAM [47], and SeqSLAM [48].”
MethodA two‐stage sampling for robust feature matching
Chou, C., Seo, Y. W., & Wang, C. (2018) · Journal of Field Robotics · doi:10.1002/rob.21778 · Cited by 5
Characterizing SLAM Benchmarks and Methods for the Robust Perception AgePreprint
Ye, W., Zhao, Y., & Vela, P. A. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.07808 · Cited by 6
“the following corpus of benchmark datasets was identified: NewCollege [1], Alderley [19], Karlsruhe [20], Ford Campus [21], Malaga 2009 [22], CMU-VL [23], TUM RGBD [16], KITTI [2], Malaga Urban [24], ICL NUIM [9], UMich NCLT [25], EuRoC [8], Nordland [26], TUM Mono [27], PennCOSYVIO [28], Zurich Urban MAV [29], RobotC…”
BackgroundSequential Place Learning: Heuristic-Free High-Performance Long-Term Place RecognitionPreprint
Chancán, M., & Milford, M. J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.02074 · Cited by 6
“In Table III, we report additional supporting comparisons against single-frame vanilla NetVLAD (pairwise) and other 6 multi-frame filtering methods (ABLE [2], ISM [42], OPR [40], VPR [41], HMM [17], SeqSLAM [30], MCN [33]) that received the pairwise similarity matrix, obtained from NetVLAD [1], for sequence filtering…”
Self-citationGlobally Consistent RGB-D SLAM With 2-D Gaussian Splatting
Zhong, X., Pan, Y., Jin, L., Popović, M., Behley, J., & Stachniss, C. (2026) · IEEE Transactions on Robotics · doi:10.1109/tro.2026.3697159 · Cited by 8
“…approaches primarily rely on aggregating handcrafted lo-cal features using bag-of-words [21], [24], random ferns [23], hamming distance embedding binary search tree [16], or VLAD [2] to build databases for efficient searching and matching [7], [43], [54], [91], or match image sequences [53], [85].”
UniPR-3D: Towards Universal Visual Place Recognition with Visual Geometry Grounded Transformer
Deng, T., Chen, X., Li, Z., Shen, H., Zhai, S., Wang, D., Civera, J., & Wang, H. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37624-4_15 · Cited by 4
A Bio-Inspired Model for Robust Navigation Assistive Devices: A Proof of Concept
Gay, S. L., Pissaloux, E. E., & Jamont, J.-P. (2023) · Communications in computer and information science · doi:10.1007/978-3-031-29548-5_2 · Cited by 4
A bio-inspired model for robust navigation assistive devices
Gay, S. L., Pissaloux, E. E., & Jamont, J.-P. (2024) · Smart Health · doi:10.1016/j.smhl.2024.100484 · Cited by 4
Automated sensory data alignment for environmental and epidermal change monitoring
Milford, M. J., Firn, J. L., Beattie, J. R., Jacobson, A., Pepperell, E., Mason, E. D., Kimlin, M. G., & Dunbabin, M. D. (2014) · USC Research Bank (University of the Sunshine Coast) · Cited by 4
Self-citationVisual Place Recognition by DTW-based sequence alignment
Hafez, A. H. A., Tello, A., & Alqaraleh, S. (2019) · Signal Processing and Communications Applications Conference (SIU) · doi:10.1109/siu.2019.8806363 · Cited by 4
“Unlike other feature-based methods, SeqSLAM uses image differences for matching.”
BackgroundMethodInfluentialRobustifying Visual Place Recognition with Semantic Scene Categorization
Arshad, S., & Kim, G.-W. (2020) · IEEE International Conference on Big Data and Smart Computing (BigComp) · doi:10.1109/bigcomp48618.2020.00-24 · Cited by 4
“Most of the existing visual SLAM approaches uses local feature detection and description methods for robust place recognition [9]–[12].”
MethodMulti-Robot Rendezvous in Unknown Environment With Limited Communication
Song, K., Chen, G., Liu, W., & Xiong, Z. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3460420 · Cited by 4
“In this paper, we employ lightweight Place Recognition (PR) descriptors [23], [24], [25] for rough RP estimation.”
MethodIndoor positioning system for warehouse environment: A scoping review
Zhang, X., Tan, Y. C., Taı, V. C., & Hao, Y. (2024) · JOURNAL OF MECHANICAL ENGINEERING AND SCIENCES · doi:10.15282/jmes.18.4.2024.9.0815 · Cited by 5
Matched Filtering Based LiDAR Place Recognition for Urban and Natural Environments
Joseph, T., Fischer, T., & Milford, M. J. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3533966 · Cited by 7
“Much of the research has focused on visual place recognition (VPR), which relies on features extracted from vision-based sensors to distinguish different locations, as demonstrated in methods like FAB-MAP [3], SeqSLAM [4], and NetVLAD [5].”
MethodSelf-citationIntrospective Loop Closure for SLAM with 4D Imaging Radar
Hilger, M., Kubelka, V., Adolfsson, D., Becker, R., Andreasson, H., & Lilienthal, A. J. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11127330 · Cited by 4
“By augmenting the query descriptor and using sequence matching [40] in viewpoint-dependent distance matrixes, they recognize places from both similar and opposing viewpoints.”
Lowry, S. M., & Milford, M. J. (2015) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2015.7353801 · Cited by 4
“The OpenSeqSLAM implementation [9] was used, but three key alterations were made.”
BackgroundMethodResultSelf-citationSimultaneous place learning and recognition for real-time appearance-based mapping
Kazmi, S. M. A. M., & Mertsching, B. (2016) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2016.7759719 · Cited by 4
“Other approaches [2], [3] localize the visual routes by means of constant or variable velocity models.”
BackgroundMethodInfluentialCondition and Viewpoint Invariant Omni-Directional Place Recognition Using CNN
Kumar, D., Neher, H., Das, A., Clausi, D. A., & Waslander, S. L. (2017) · Canadian Conference on Computer and Robot Vision · doi:10.1109/crv.2017.26 · Cited by 4
“RELATED WORK In the past, numerous place recognition methods have been proposed for large scale environment localization, including FAB MAP [11], [23], MonoSLAM [24] and SeqSLAM [20], but, there are very few long term localization place recognition systems [5], [25].”
BackgroundMethodInfluentialLearning-based scene recognition with monocular camera for light-rail system
Yao, M., Siu, W.-C., & Jia, K. (2018) · IEEE International Conference on Industrial Electronics for Sustainable Energy Systems (IESES) · doi:10.1109/ieses.2018.8349879 · Cited by 4
“When we select more matched frames, the SeqSLAM method can reach the highest precision of 95.50% (with the best four frames are chosen).”
MethodInfluentialTensorMap: LIDAR-BASED TOPOLOGICAL MAPPING AND LOCALIZATION VIA TENSOR DECOMPOSITIONS
Rambhatla, S., Sidiropoulos, N. D., & Haupt, J. (2018) · IEEE Global Conference on Signal and Information Processing · doi:10.1109/globalsip.2018.8646665 · Cited by 4
“Prior-Art Building topological maps for localization using imaging-based techniques has gained traction in recent times since these are inexpensive to implement and faster to process [2–9], as compared to Lidar sensors.”
MethodBTEL: A Binary Tree Encoding Approach for Visual Localization
Le, H., Hoang, T., & Milford, M. J. (2019) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2019.2932580 · Cited by 5
“niques [3] make use of visual similarity between image descriptors to assist the localization process.”
BackgroundMethodSelf-citationAn efficient edge-feature constraint visual SLAM
Liu, W., Mo, Y., & Jiao, J. (2019) · International Conference on Artificial Intelligence, Information Processing and Cloud Computing (AIIPCC) · doi:10.1145/3371425.3371455 · Cited by 4
Improved Deep Distance Learning for Visual Loop Closure Detection in Smart City
Jin, S., Gao, Y. H., & Chen, L. (2020) · Peer-to-Peer Networking and Applications · doi:10.1007/s12083-019-00861-w · Cited by 4
“Hence, Milford et al. [9] proposed a post-processing method by comparing the similarity be-tween images based on the sum of absolute differences.”
BackgroundMethodSPRINT: Subgraph Place Recognition for INtelligent Transportation
Latif, Y., Doan, A.-D., Chin, T.-J., & Reid, I. D. (2020) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra40945.2020.9196522 · Cited by 4
“Place recognition methods can take advantage of the temporal relations between images [14, 8] to address appearance changes over the life time of a robot.”
BackgroundA sequence-based visual place recognition method for aerial mobile robots
Yang, M., Mao, J., He, X., Zhang, L., & Hu, X. (2020) · Journal of Physics Conference Series · doi:10.1088/1742-6596/1654/1/012080 · Cited by 4
“However, the original SeqSLAM requires the constant velocity of vehicle and exhaustive sequence match searching through multiple locations and image scales.”
BackgroundMethodInfluentialReview of Place Recognition Approaches: Traditional and Deep Learning Methods
Taha, M. A., Şah, M., & Direkoğlu, C. (2021) · Advances in intelligent systems and computing · doi:10.1007/978-3-030-64058-3_22 · Cited by 4
Keyframes retrieval for robust long-term visual localization in changing conditions
Bouaziz, Y., Royer, E., Bresson, G., & Dhome, M. (2021) · IEEE World Symposium on Applied Machine Intelligence and Informatics (SAMI) · doi:10.1109/sami50585.2021.9378614 · Cited by 4
“Milford and Wyeth [5] proposed to enhance the performance of global image descriptors for places recognition by matching sequences of images («SeqSLAM») and they achieved good results on various seasonal datasets.”
BackgroundInterval Inspired Approach Based on Temporal Sequence Constraints to Place Recognition
Neuland, R., Rodrigues, F., Pittol, D., Jaulin, L., Maffei, R. D. Q., Kolberg, M., & Prestes, E. (2021) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-021-01375-5 · Cited by 4
“The use of vision as the central sensing modality has shown exciting results for learning and recognizing places [21].”
BackgroundMethodInfluentialA First Look at Private Communications in Video Games using Visual Features
Wajid, A., Kamal, N., Sharjeel, M., Sheikh, R. M., Wasim, H. B., Ali, M. H., Hussain, W., Ali, S. T., & Anjum, L. (2021) · Proceedings on Privacy Enhancing Technologies · doi:10.2478/popets-2021-0055 · Cited by 4
“Prior work demonstrates that enhancements in image contrast provide light invariance, enabling comparable performance for scenes with different lighting conditions [37], as typically observed during the day/night cycle [50] and seasonal changes [56].”
BackgroundDeep Next-Best-View Planner for Cross-Season Visual Route Classification
Kurauchi, K., & Kanji, T. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9412043 · Cited by 5
“They are largely divided into global [15] and local feature approaches [16].”
BackgroundSequence-based visual place recognition: a scale-space approach for boundary detection
Bampis, L., & Gasteratos, A. C. (2021) · Autonomous Robots · doi:10.1007/s10514-021-09984-7 · Cited by 4
“Nevertheless, the dynamic segmentation of image sequences based on the proposed Scale-Space analysis is capable of considerably improving the performance for both SeqSLAM and the already competitive approach of Bampis et al. (2016).”
BackgroundMethodInfluentialLearning-based Camera Relocalization with Domain Adaptation via Image-to-Image Translation
Yang, C., Liu, Y., & Zell, A. (2021) · International Conference on Unmanned Aircraft Systems (ICUAS) · doi:10.1109/icuas51884.2021.9476673 · Cited by 4
Visual Place Recognition via Local Affine Preserving Matching
Ye, X., & Ma, J. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9561709 · Cited by 5
“Figure 8 compares the AUC, MR and MP results of our VPR method with SeqSLAM [34], AlexNet [14], Region-BoW [35] and Region-VLAD [16].”
MethodBinary Graph Descriptor for Robust Relocalization on Heterogeneous Data
Wang, X., Christie, M., & Marchand, É. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3142854 · Cited by 4
Bouaziz, Y., Royer, E., Bresson, G., & Dhome, M. (2022) · Multimedia Tools and Applications · doi:10.1007/s11042-021-11870-4 · Cited by 4
“Milford and Wyeth [35] proposed to enhance the performance of global image descriptors for places recognition by matching sequences”
BackgroundChen, Y., Yang, Z. X., Wang, W., & Peng, H. (2022) · Journal of Electronic Imaging · doi:10.1117/1.jei.31.6.061816 · Cited by 4
Binary Image Fingerprint: Stable Structure Identifier for 3D LiDAR Place Recognition
Zhang, G., Zhang, T., zhao, S., & Hou, L. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3297063 · Cited by 5
“FAB-MAP [1], SeqSLAM [2] and ABLE-M [3] techniques also show excellent performance in long-term visual localization.”
BackgroundMethodLocalization Exploiting Semantic and Metric Information in Non-static Indoor Environments
Gómez, C., Hernandez, A. C., Barber, R., & Stachniss, C. (2023) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-023-02021-y · Cited by 4
“Localization in outdoor dynamic environments has been approached from multiple perspectives: appearance-based methods that exploit image sequences [19, 20] or navigation sequences [21] and methods that exploit semantics [22–24].”
AdaptSeqVPR: An Adaptive Sequence-Based Visual Place Recognition Pipeline
Li, H., Peng, G., Zhang, J., Vaikundam, S., & Wang, D. (2023) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros55552.2023.10341533 · Cited by 5
“In this case, using a predefined uniform velocity to retrieve sequences as in SeqSLAM [6] cannot accurately fit the true trajectory.”
BackgroundMethodInfluentialVisual impression localization of autonomous robots
Boubou, S., Hafez, A. H. A., & Suzuki, E. (2015) · IEEE International Conference on Automation Science and Engineering (CASE) · doi:10.1109/coase.2015.7294100 · Cited by 4
“While landmark based methods require extractions and reassignments of distinct visual landmarks, appearance-based methods use a description of the view at a certain point [3], [4], [5], [6], leading to a more global impression of the scene.”
MethodUnsupervised Feature Learning for Visual Place Recognition in Changing Environments
Zhao, D., Si, B., & Tang, F. (2019) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2019.8852466 · Cited by 4
“Although SeqSLAM has some successful applications in changing environments, extensive research still need to be carried out on learning effective features for dynamic environments.”
BackgroundMethodInfluentialYusuke, K., Kanji, T., Naiming, Y., Yuji, H., & Yamaguchi, K. (2019) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc.2019.8917170 · Cited by 4
“However, SLAM in non-stationary environments remains a significant challenge [2]–[5].”
BackgroundBio-inspired Multi-scale Visual Place Recognition for the Aerial Vehicle Navigation
Fan, C., Jacobson, A., Chen, Z., He, X., Zhang, L., Hu, X., & Milford, M. J. (2021) · Lecture notes in electrical engineering · doi:10.1007/978-981-15-8155-7_87 · Cited by 4
Self-citationTowards CNN Map Compression for camera relocalisationPreprint
Contreras, L., & Mayol-Cuevas, W. W. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.00845 · Cited by 4
“[27] defines a map as a collection of images and uses image batch matching to find the current position in the map.”
BackgroundLoop Closure Detection with RGB-D Feature Pyramid Siamese NetworksPreprint
Zhang, Q., Mai, A., Menke, J. T., & Yang, A. Y. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1811.09938 · Cited by 4
“In addition, these image representations describe the local appearance of individual patches, limiting their descriptive power with respect to global descriptor methods [37, 19].”
BackgroundFaVoR: Features via Voxel Rendering for Camera Relocalization
Polizzi, V., Cannici, M., Scaramuzza, D., & Kelly, J. W. (2025) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61041.2025.00015 · Cited by 3
“Sequence-based place recognition, for instance, exploits consistency across frames to identify previously visited locations [27], while geometric constraints derived from point cloud alignment and 3D-3D matching can provide additional, useful information [23,24,47].”
BdFusion: Bi-directional visual-LiDAR fusion for resilient place recognition
Liang, A., Chen, Z., Xiong, W., Meng, F., Yin, Y., Zhang, D., & Li, Q. (2025) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2025.07.022 · Cited by 4
Improving Visual Place Recognition with Sequence-Matching Receptiveness Prediction
Hussaini, S., Fischer, T., & Milford, M. J. (2025) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros60139.2025.11247581 · Cited by 3
“One of the common methods for improving the performance of VPR systems in challenging conditions is to filter input over time and space, for example, through sequence matching techniques [11]–[17].”
Self-citationFrom Gaming to Research: GTA V for Synthetic Data Generation for Robotics and Navigations
Scucchia, M., Arranz, P., Ferrara, M., & Maltoni, D. (2025) · International Conference on Robotics and Computer Vision (ICRCV) · doi:10.1109/icrcv67407.2025.11349201 · Cited by 7
“For these experiments, we used three datasets: • Alderley [12], a real dataset of paired day/night RGB images captured by a camera on top of a car in an urban environment.”
Co-VisiON: Co-Visibility ReasONing on Sparse Image Sets of Indoor Scenes
Chen, C., Dang, N., Zhang, J., Sun, W., Zheng, P., He, X., Ye, Y., Zhang, J., Srinivas, T., & Feng, C. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00504 · Cited by 3
The RGB-D Triathlon: Towards Agile Visual Toolboxes for Robots
Cermelli, F., Mancini, M., Ricci, E., & Caputo, B. (2019) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros40897.2019.8968562 · Cited by 3
Long-term Visual Place Recognition
Alijani, F., Peltomäki, J., Puura, J., Huttunen, H. J., Kämäräinen, J., & Rahtu, E. (2022) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr56361.2022.9956392 · Cited by 4
“Bai et al. [12] proposed SeqCNNSLAM to improve the viewpoint invariance and boost the performance of the original SeqSLAM [4].”
A Hierarchical Utilization of Semantic Gradients and Scene Structure for Visual Place Recognition
Bao, Y., Pan, Y., Yang, Z., & Huan, R. (2023) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2023.3281870 · Cited by 3
“For VPR, it is a common practice to utilize temporal sequence information to improve the performance, especially under challenging conditions [25], [41], [42].”
BackgroundMethodSelf-Supervised Learning for Place Representation Generalization across Appearance Changes
Musallam, M. A., Gaudillière, V., & Aouada, D. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00728 · Cited by 8
Visual place recognition from end-to-end semantic scene text features
Raisi, Z., & Zelek, J. (2024) · Frontiers in Robotics and AI · doi:10.3389/frobt.2024.1424883 · Cited by 3
“Table 1 presents the results of the quantitative evaluation comparing our proposed model with several state-of-the-art methods (Cummins and Newman, 2008; Milford and Wyeth, 2012; Arandjelovic et al., 2016; Anoosheh et al., 2019; Hong et al., 2019) on the SCTP dataset (Hong et al., 2019).”
MethodAdaptive Thresholding for Sequence-Based Place Recognition
Vysotska, O., Bogoslavskyi, I., Hutter, M., & Stachniss, C. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11128422 · Cited by 3
“Milford and Wyeth [15] propose to use a similar concept but in the context of alignment of two image sequences that were recorded in different period in time rather than for loop closure candidate selection in SLAM.”
EVLoc: Event-Based Visual Localization in LiDAR Maps via Event-Depth Registration
Chen, K., Zhang, J., & Fraundorfer, F. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11127535 · Cited by 3
“Finally, they use the common-used SeqSLAM algo-rithm [18] to perform place recognition.”
Learning Long-Term Invariant Features for Vision-Based Localization
Mithun, N. C., Simons, C., Casey, R. T., Hilligardt, S., & Roy–Chowdhury, A. K. (2018) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2018.00225 · Cited by 3
“On the contrary, global image descriptors, like GIST suffer from high sensitivity to viewpoint change [31, 35].”
BackgroundAvoiding to Face the Challenges of Visual Place Recognition
Mihankhah, E., & Wang, D. (2018) · Advances in intelligent systems and computing · doi:10.1007/978-3-030-01054-6_52 · Cited by 3
“One other suggestion would be to study places in sequences similar to the approaches suggested by [67] and [68].”
c-M2DP: A Fast Point Cloud Descriptor with Color Information to Perform Loop Closure Detection
Perdomo, L., Pittol, D., Mantelli, M., Maffei, R. D. Q., Kolberg, M., & Prestes, E. (2019) · IEEE International Conference on Automation Science and Engineering (CASE) · doi:10.1109/coase.2019.8842896 · Cited by 3
Three level sequence-based Loop Closure Detection
Rodrigues, F., Neuland, R., Mantelli, M., Pittol, D., Maffei, R. D. Q., Prestes, E., & Kolberg, M. (2020) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2020.103620 · Cited by 3
Traffic Intersection Re-Identification Using Monocular Camera Sensors
Xiong, L., Deng, Z., Huang, Y., Du, W., Zhao, X., Lu, C., & Tian, W. (2020) · Sensors · doi:10.3390/s20226515 · Cited by 3
“One of the related datasets is the Alderley Day/Night Dataset [27], which was captured along a fixed route in two different conditions: sunny daytime and rainy nighttime.”
BackgroundDynamic Time Warping of Deep Features for Place Recognition in Visually Varying Conditions
Alqaraleh, S., Hafez, A. H. A., & Tello, A. (2021) · Arabian Journal for Science and Engineering · doi:10.1007/s13369-020-05146-6 · Cited by 3
“, Holistic [22], Only look once [27], NetVLAD [19] and SeqSLAM [15].”
BackgroundEgomap: Hierarchical First-Person Semantic Mapping
Süveges, T., & McKenna, S. J. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-68796-0_25 · Cited by 3
“SeqSLAM compares sequences of observations to detect loop closures [14].”
BackgroundHafez, A. H. A., Tello, A., & Alqaraleh, S. (2021) · Arabian Journal for Science and Engineering · doi:10.1007/s13369-021-06148-8 · Cited by 3
VGG-CAE: Unsupervised Visual Place Recognition Using VGG16-Based Convolutional Autoencoder
Xu, Z., Zhang, Q., Hao, F., Ren, Z., Kang, Y., & Cheng, J. S. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-88007-1_8 · Cited by 4
Visual Place Recognition via Semantic and Geometric Descriptor for Automated Valet Parking
Yu, J., & Su, J. (2021) · IEEE International Conference on Robotics and Biomimetics (ROBIO) · doi:10.1109/robio54168.2021.9739413 · Cited by 3
“Three state-of-the-art methods are chosen as the baseline, namely ORB feature with the bag-of-words model (DBoW) [3], Seq [5] and LoST [13].”
MethodInfluentialSEER: Unsupervised and sample-efficient environment specialization of image descriptors
Neubert, P., & Schubert, S. (2022) · Robotics: Science and Systems · doi:10.15607/rss.2022.xviii.006 · Cited by 3
“To integrate these image retrieval methods into the robotics context of place recognition, there is a wide range of methods, e.g. [26, 35, 29], that exploit the structure of the robotics task or additional available information about the database and query set for performance improvements.”
Method
Metadata
- Authors
- Milford, M. J., & Wyeth, G. F.
- Affiliations
- Queensland University of Technology (AU)
- Venue
- 2012 IEEE International Conference on Robotics and Automation
- Publisher
- IEEE
- Published
- 2012-05
- Pages
- 1643-1649
- ISSN
- 1050-4729 (ISSN-L)
- OpenAlex
- W2110405746
- Cited by
- 981
- References
- 22
- Citations by year
- 2026: 22 · 2025: 44 · 2024: 66 · 2023: 61 · 2022: 82 · 2021: 115 · 2020: 110 · 2019: 125 · 2018: 99 · 2017: 74 · 2016: 52 · 2015: 61 · 2014: 38 · 2013: 25 · 2012: 7
- FWCI
- 78.04
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- Proceedings - IEEE International Conference on Robotics and Automation/Proceedings (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Robotics and Sensor-Based Localization
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Keywords
- visual SLAM, visual navigation, route recognition, seasonal variation, heavy rainfall, vehicle-mounted camera, trajectory matching
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.