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 401–500 of 987
Junior, L. J. L., Oliveira-Santos, T., Badue, C. S., & Souza, A. F. D. (2015) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2015.7139699 · Cited by 14
“Directions for future work include to address the issues raised above, and to extend the VIBML to perform localization in widely used image-maps like Google Street View.”
MethodVisual localization in highly crowded urban environments
Hafez, A. H. A., Singh, M., Krishna, K. M., & Jawahar, C. V. (2013) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2013.6696749 · Cited by 14
“Localization in outdoor environments is challenging because, queries are often from different viewpoint, scale and illumination than the previous visual experience [3].”
BackgroundMethodCho, S. W., Baek, N. R., Koo, J. H., & Park, K. R. (2020) · IEEE Access · doi:10.1109/access.2020.3048366 · Cited by 14
“The road scene open databases acquired in a nighttime environment include Alderley [33], Berkeley deep drive 100K (BDD100K) [34], and Nighttime Driving [11] database.”
BackgroundSpatio-Temporal Memory for Navigation in a Mushroom Body Model
Zhu, L., Mangan, M., & Webb, B. H. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-64313-3_39 · Cited by 16
Fusion of visual odometry and place recognition for SLAM in extreme conditions
Yadav, R., & Kala, R. (2022) · Applied Intelligence · doi:10.1007/s10489-021-03050-6 · Cited by 15
“• SeqSLAM [21]: SeqSlam (available at https:// github.”
MethodZhang, Q., Xu, Z., Kang, Y., Hao, F., Ren, Z., & Cheng, J. S. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.111015 · Cited by 14
Lobefaro, L., Malladi, M. V. R., Vysotska, O., Guadagnino, T., & Stachniss, C. (2023) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros55552.2023.10342449 · Cited by 14
“Milford et al. [18] propose to exploit sequential information of the input data for VPR in autonomous driving and transfer the single image matching into a sequence matching problem using multiple images.”
BackgroundImproving robustness of monocular urban localization using augmented Street View
Yu, L., Joly, C., Bresson, G., & Moutarde, F. (2016) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc.2016.7795603 · Cited by 14
“A series of methods has been proposed regarding to these three aspects in the literature, like FABMAP [4] and SeqSLAM [5].”
MethodVisual Place Recognition in Long-term and Large-scale Environment based on CNN Feature
Zhu, J., Ai, Y., Tian, B., Cao, D., & Scherer, S. (2018) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2018.8500686 · Cited by 14
“Then, we compare the performance of our method with the main state-of-the-art works, include FAB-MAP, SeqSLAM and ABLE-M algorithms.”
MethodInfluentialRobust simultaneous localization and mapping in low‐light environment
Huang, J., & Liu, S. (2019) · Computer Animation and Virtual Worlds · doi:10.1002/cav.1895 · Cited by 17
“However, a feature method at the time is still not good enough to meet the needs of a SLAM system, which should have the ability to recognize the same scene in different seasons or both day and night, so Milford et al. 10 proposed a SLAM system based on an image sequence.”
MethodLoop closure detection with patch-level local features and visual saliency prediction
Jin, S., Dai, X.-Y., & Meng, Q. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.105902 · Cited by 14
(2018) · Digital Repository (National Repository of Grey Literature) · Cited by 13
VLAD-BuFF: Burst-Aware Fast Feature Aggregation for Visual Place Recognition
Khaliq, A., Xu, M., Hausler, S., Milford, M. J., & Garg, S. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72784-9_25 · Cited by 13
“…fronts: backbone networks [7,31,70], objective functions [8,51,62,77], multi-scale processing [25,74,77], semantics [21–23,48 ], sequences [20 ,26 ,43 NetVLAD [3] and its follow-up VLAD enhancements in VPR [25, 28, 31–33, 35, 40, 41, 77], including the most recent work SALAD [28], are all reliant…”
BackgroundSelf-citationDetection-by-Localization: Maintenance-Free Change Object Detector
Kanji, T. (2019) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2019.8793482 · Cited by 13
“However, the BoVF approach suffers from appearance variations in places like seasonal changes and cyclic day-night illumination changes [52].”
BackgroundShu, M., Chen, G., & Zhang, Z. (2021) · IEEE Access · doi:10.1109/access.2021.3060760 · Cited by 14
Self-Supervised Feature Learning for Long-Term Metric Visual Localization
Chen, Y., & Barfoot, T. D. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3227866 · Cited by 14
“For our purposes, we adopt SeqSLAM aided by Visual Odometry (VO) to find direct image correspondences between temporally neighbouring experiences, and indirectly link further experiences by building an offline experience-association graph.”
BackgroundMethodInfluentialUnsupervised online learning of condition-invariant images for place recognition
Lowry, S. M., Wyeth, G. F., & Milford, M. J. (2014) · QUT ePrints (Queensland University of Technology) · Cited by 13
Self-citation2D visual place recognition for domestic service robots at night
Mount, J., & Milford, M. J. (2016) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2016.7487686 · Cited by 13
“A multitude of mature vision-based SLAM systems are available including MonoSLAM [2], FrameSLAM [6], VGPS [7], Mini-SLAM [8] and several others [9-16].”
MethodSelf-citationA Bio-Inspired Goal-Directed Visual Navigation Model for Aerial Mobile Robots
Mao, J., Hu, X., Zhang, L., He, X., & Milford, M. J. (2020) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-020-01190-4 · Cited by 14
“Such topological information has been used to augment the place recognition performance in a robot ’ s localization system [40, 41].”
MethodSelf-citationBiologically-inspired visual place recognition with adaptive multiple scales
Fan, C., Chen, Z., Jacobson, A., Hu, X., & Milford, M. J. (2017) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2017.07.015 · Cited by 13
“On the St. Lucia dataset, when using Grayscale and GIST features, the AMS system consistently outperforms FABMAP [10] and SeqSLAM [18] algorithm.”
MethodResultInfluentialSelf-citationAccurate Vision-based Vehicle Localization using Satellite ImageryPreprint
Chu, H., Mei, H., Bansal, M., & Walter, M. R. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1510.09171 · Cited by 15
“[40] seek to overcome the brittleness of pointbased features to environment variation [7, 10, 43, 44] by learning a region-based detector to improve invariance to appearance changes.”
BackgroundAutomatic image scaling for place recognition in changing environments
Pepperell, E., Corke, P., & Milford, M. J. (2015) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2015.7139316 · Cited by 12
“Precision refers to the proportion of returned frame pairs that were correct (within 40 m of ground truth – similar to past studies [5], [24]; though the majority were within 20 m), and recall is the proportion of total correct frame pairs that were returned.”
BackgroundMethodResultInfluentialSelf-citationMining visual phrases for long-term visual SLAM
Kanji, T., Yuuto, C., & Masatoshi, A. (2014) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2014.6942552 · Cited by 13
“[26] investigated the robustness of the place recognition of the SeqSLAM algorithms in changing environments across all four seasons on a 3000 km journey.”
BackgroundInfluentialVisual Odometry and Place Recognition Fusion for Vehicle Position Tracking in Urban Environments
Ouerghi, S., Boutteau, R., Savatier, X., & Tlili, F. (2018) · Sensors · doi:10.3390/s18040939 · Cited by 12
“Significant performance improvements over Fab-MAP under extreme lighting and atmospheric variations have been achieved within the SeqSLAM system, which relies on the use of a whole image regardless of its content by searching in a pre-learnt database of images for the most similar sequence to the query sequence [1,2].”
Scale-Robust Localization Using General Object Landmarks
Holliday, A. G. C., & Dudek, G. (2018) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594011 · Cited by 12
“In a similar fashion, Vysotska et al. [28] use whole-image descriptors from a CNN in a SeqSLAMlike framework.”
MethodVisual Place Recognition via Robust ℓ2-Norm Distance Based Holism and Landmark Integration
Liu, K., Wang, H., Han, F., & Zhang, H. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33018034 · Cited by 12
“Baseline and recent methods, including the BRIEF-GIST (Sünderhauf and Protzel 2011), Normalized Gradients (NormG) of grayscale images (used in SeqSLAM (Milford and Wyeth 2012)), and the methods that simply based upon color, LBP, SURF, CNN, HOG features, are compared in our experiments.”
MethodLearning Integrated Holism-Landmark Representations for Long-Term Loop Closure Detection
Han, F., Wang, H., & Zhang, H. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.12101 · Cited by 12
“In addition, several baseline and recent methods are compared in each experiment, including the BRIEF-GIST (Sünderhauf and Protzel 2011), Normalized Gradients (NormG) of grayscale images (used in SeqSLAM (Milford and Wyeth 2012)), and techniques based only upon color, LBP, HOG, SURF, CNN features.”
MethodVisual Global Localization with a Hybrid WNN-CNN Approach
Forechi, A., Oliveira-Santos, T., Badue, C. S., & Souza, A. F. D. (2018) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2018.8489335 · Cited by 12
Towards View-Invariant Intersection Recognition from Videos using Deep Network Ensembles
Kumar, A., Gupta, G., Sharma, A. K., & Krishna, K. M. (2018) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594449 · Cited by 12
“Real-world datasets such as [17], [18] are designed with the idea of visual place recognition using images.”
BackgroundLi, C.-T., & Siu, W.-C. (2020) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2020.2975710 · Cited by 12
“Reference [21] assumed constant speed situations which are not necessarily true for reality.”
BackgroundMethodInfluentialTowards a Predictive Bio-Inspired Navigation Model
Gay, S. L., Run, K. L., Pissaloux, E. E., Romeo, K., & Lecomte, C. (2021) · Information · doi:10.3390/info12030100 · Cited by 12
Seifi, F., Azizi, M. J., & Niaki, S. T. A. (2021) · Computers & Industrial Engineering · doi:10.1016/j.cie.2021.107581 · Cited by 12
Underwater robot visual place recognition in the presence of dramatic appearance change
Li, J., Eustice, R. M., & Johnson‐Roberson, M. (2015) · MTS/IEEE OCEANS Conference (OCEANS) · doi:10.23919/oceans.2015.7404369 · Cited by 12
“SeqSLAM [3] localizes the robot using a topological strategy that jointly considers visual feature comparisons along trajectory segments to increase the robustness of dramatic scene changes in contrast to individual images.”
MethodBeyond a Shadow of a Doubt: Place Recognition with Colour-Constant Images
MacTavish, K., Paton, M., & Barfoot, T. D. (2016) · Springer tracts in advanced robotics · doi:10.1007/978-3-319-27702-8_13 · Cited by 12
“[20] use Sequence SLAM (SeqSLAM) [14] with monochromatic images to localize a train over 3000 km across seasons with impressive results, but do not perform full Simultaneous Localization and Mapping (SLAM) with the ability to add new places.”
MethodJacobson, A., Chen, Z., & Milford, M. J. (2018) · Biological Cybernetics · doi:10.1007/s00422-017-0745-7 · Cited by 13
“9: Precision-Recall curves for the night-time dataset using SeqSLAM with a sequence length 20.”
BackgroundMethodInfluentialSelf-citationYu, L., Jacobson, A., & Milford, M. J. (2018) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2018.2792144 · Cited by 13
“Alternatively, we could apply some existing matching schemes such as SeqSLAM [26], to further improve the stability by taking consideration of multi-frame integration.”
MethodSelf-citationJIST: Joint Image and Sequence Training for Sequential Visual Place Recognition
Berton, G., Trivigno, G., Barbara Caputo, B., & Masone, C. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3339058 · Cited by 11
“However, we want to remark that in principle our multi-task framework is loss-agnostic, so the im2im loss can be easily swapped with another one, for example if a more performing loss becomes available.”
BackgroundZou, Z., Zheng, C., Yuan, C., Zhou, S., Xue, K., & Zhang, F. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3483032 · Cited by 11
“One category of placerecognitionmethodsreliesonimagescapturedbycameras, exemplifiedbyFAB-MAP[3],DBoW2[2],andSeqSLAM[4].”
Towards Brain-based Sensor Fusion for Navigating Robots
Jacobson, A., & Milford, M. J. (2012) · QUT ePrints (Queensland University of Technology) · Cited by 11
Self-citationImpact of Next-Generation Mobile Technologies on IoT-Cloud Convergence
Hossain, M. S., Xu, C., Li, Y., Pathan, A. K., Bilbao, J., Zeng, W., & Saddik, A. E. (2017) · IEEE Communications Magazine · doi:10.1109/mcom.2017.7823332 · Cited by 12
Park, C.-S., Chae, H.-W., & Song, J.-B. (2020) · International Journal of Control Automation and Systems · doi:10.1007/s12555-019-0891-x · Cited by 11
“SeqSLAM [7], which uses sequence information of an image, has been proposed.”
MethodInsect Inspired View Based Navigation Exploiting Temporal Information
Kagioulis, E., Philippides, A. O., Graham, P., Knight, J. C., & Nowotny, T. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-64313-3_20 · Cited by 12
“One possibility is that views could be stored as a sequence, and whilst this could be problematic [25, 21], sequence information has been shown to be useful in engineering approaches to navigation [17, 5].”
BackgroundImproving Visual Place Recognition Performance by Maximising Complementarity
Waheed, M., Milford, M. J., McDonald-Maier, K. D., & Ehsan, S. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3088779 · Cited by 17
Self-citationQin, C., Zhang, Y., Liu, Y., Coleman, S., Du, H., & Kerr, D. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.06.038 · Cited by 11
“The proposed approach significantly outperforms all other methods except for the SeqSLAM.”
BackgroundMethodInfluentialHussaini, S., Milford, M. J., & Fischer, T. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160749 · Cited by 15
“[14] and conventional methods NetVLAD [28], DenseVLAD [29] and SAD [30].”
MethodInfluentialSelf-citationVPRTempo: A Fast Temporally Encoded Spiking Neural Network for Visual Place Recognition
Hines, A., Stratton, P. G., Milford, M. J., & Fischer, T. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10610918 · Cited by 17
“Several systems have been deployed to tackle this problem, ranging from simple approaches such as Sum of Absolute Differences (SAD) [33], over handcrafted systems like DenseVLAD [13] to learned approaches like NetVLAD [11], [33] and many others including [12], [14], [15].”
BackgroundMethodInfluentialSelf-citationGSPR: Multimodal Place Recognition Using 3D Gaussian Splatting for Autonomous Driving
Qi, Z., Ma, J., Xu, J., Zhou, Z., Cheng, L., & Xiong, G. (2025) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros60139.2025.11247374 · Cited by 11
“Place recognition is a classic topic in the fields of robotics and computer vision, and there have been various types of traditional methods based on handcrafted descriptors [10], [1], [11].”
Vision-based Markov localization across large perceptual changes
Naseer, T., Suger, B., Ruhnke, M., & Burgard, W. (2015) · European Conference on Mobile Robots · doi:10.1109/ecmr.2015.7324181 · Cited by 11
Online place recognition calibration for out-of-the-box SLAM
Jacobson, A., Chen, Z., & Milford, M. J. (2015) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2015.7353544 · Cited by 11
“We specifically select the SAD or L(1)-norm comparison metric as the place recognition method for this study, opposed to other place matching techniques [1], [22], [23], as SAD enables place recognition performance to be evaluated by varying a single parameter with no specific environmental information or vocabulary l…”
MethodSelf-citationSparse optimization for robust and efficient loop closing
Latif, Y., Huang, G. P., Leonard, J. J., & Neira, J. (2017) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2017.03.016 · Cited by 11
“Note that if the information invariant to such changes can be extracted as in [13, 21, 33, 12, 36, 28, 29], the proposed formulation can also be used to obtain loop-closure hypotheses.”
BackgroundMethodAppearance-invariant place recognition by adversarially learning disentangled representation
Qin, C., Zhang, Y., Liu, Y., Coleman, S., Kerr, D., & Lv, G. (2020) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2020.103561 · Cited by 11
“The Alderley dataset was first introduced in SeqSLAM [38].”
Fang, Y., Wang, K., Cheng, R., & Yang, K. (2020) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv47402.2020.9304612 · Cited by 11
“Similarly, SeqSLAM [25], as a holistic algorithm based on sequence matching, is also found not robust to viewpoint changes, although it performs remarkably well under appearance changes.”
BackgroundQin, C., Zhang, Y., Liu, Y., Zhu, D., Coleman, S., & Kerr, D. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2021.3105175 · Cited by 12
Multiple Maps for the Feature-based Monocular SLAM System
Hu, H., Sun, H., Ye, P., Jia, Q., & Gao, X. (2018) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-018-0829-1 · Cited by 11
“Another typical example is SeqSLAM [10] presented by Milford and Wyeth, which matches the current image with a sequences of images.”
MethodLearning Place-and-Time-Dependent Binary Descriptors for Long-Term Visual Localization
Zhang, N., Warren, M., & Barfoot, T. D. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8460674 · Cited by 11
“Milford et al. [13] proposed the use a sequence of images to localize rather a single one.”
BackgroundMaldonado-Ramírez, A., & Torres-Méndez, L. A. (2019) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2019.8793550 · Cited by 11
“We have obtained good results in terms of precision and recall compared to other approaches, such as SeqSLAM [39] or FAB-MAP [35].”
ResultZhao, C., Ding, R., & Key, H. L. (2019) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2019.8802931 · Cited by 11
“Moreover, compared with SeqSLAM which is a traditional method, both our method and other deep learning based methods have a significant progress, which prove that deep learning based features perform much better than hand-crafted features on this problem.”
MethodInfluentialFlow-based SLAM: From geometry computation to learning
Yan, Z., & Zha, H. (2019) · Virtual Reality & Intelligent Hardware · doi:10.1016/j.vrih.2019.09.001 · Cited by 12
Jiang, M., Feng, X., Song, S., Herrmann, J. M., Li, J., Li, Y., Hu, Z., Li, Z., Liu, J., & Li, S. (2019) · IEEE Access · doi:10.1109/access.2019.2952445 · Cited by 13
“[24] M. J. Milford and G. F. Wyeth, ‘‘SeqSLAM: Visual route-based navigation for sunny summer days and stormy winter nights,’’ in Proc.”
BackgroundTowards CNN Map Representation and Compression for Camera Relocalisation
Contreras, L., & Mayol-Cuevas, W. W. (2018) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2018.00067 · Cited by 12
Place Recognition with Event-based Cameras and a Neural Implementation of SeqSLAMPreprint
Milford, M. J., Kim, H., Mangan, M., Leutenegger, S., Stone, T., Webb, B. H., & Davison, A. J. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1505.04548 · Cited by 11
Self-citationKhaliq, A., Ehsan, S., Milford, M. J., & McDonald-Maier, K. D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.08153 · Cited by 11
Self-citationOn Motion Blur and Deblurring in Visual Place Recognition
Ismagilov, T., Ferrarini, B., Milford, M. J., Tuyen, N. T. V., Ramchurn, S. D., & Ehsan, S. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3554103 · Cited by 10
“The impact of these changes on VPR performance has been extensively investigated [1], [2], [3], through numerous datasets available in the literature [4].”
Self-citationPlace Recognition Based Visual Localization Using LBP Feature and SVM
Qiao, Y., Cappelle, C., & Ruichek, Y. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-27101-9_30 · Cited by 10
Localization on freeways using the horizon line signature
Ho, N., & Chakravarty, J. (2014) · Lirias · Cited by 10
Synchronous Adversarial Feature Learning for LiDAR based Loop Closure Detection
Yin, P., He, Y., Xu, L., Peng, Y., Han, J., & Xu, W. (2018) · Proceedings of the ... American Control Conference/Proceedings of the American Control Conference · doi:10.23919/acc.2018.8431776 · Cited by 10
Schubert, S., Neubert, P., & Protzel, P. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3052446 · Cited by 10
EchoVPR: Echo State Networks for Visual Place Recognition
Özdemir, A., Scerri, M., Barron, A. B., Philippides, A. O., Mangan, M., Vasilaki, E., & Manneschi, L. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3150505 · Cited by 11
“Milford and Wyeth [11] were the first to demonstrate that such issues could be overcome by matching sequences of images using a global search to overcome individual image mismatches.”
BackgroundMethodVisual Place Recognition for Opposite Viewpoints and Environment Changes
Li, P., Wen, S., Xu, C., & Qiu, T. Z. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3350152 · Cited by 10
“Over the past decade, numerous studies have explored VPR with sequence information [30], [31], [32], [33], [34], [35].”
BackgroundLGD: A fast place recognition method based on the fusion of local and global descriptors
Wang, W., Min, H., Wu, X., Yang, L., Yan, C., Fang, Y., & Zhao, X. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.123996 · Cited by 10
In the Dead of Winter: Challenging Vision-Based Path Following in Extreme Conditions
Paton, M., Pomerleau, F., & Barfoot, T. D. (2016) · Springer tracts in advanced robotics · doi:10.1007/978-3-319-27702-8_37 · Cited by 13
“Proposed solutions for localization across appearance change either provide only topological localization [10, 11], require offline collection of the scene in multiple appearances [2, 9], or have under-performed in winter environments [14].”
BackgroundSemantic Monte-Carlo localization in changing environments using RGB-D cameras
Himstedt, M., & Maehle, E. (2017) · European Conference on Mobile Robots · doi:10.1109/ecmr.2017.8098711 · Cited by 10
“suggested to identify place correspondences across substantive changes by utilizing sequences of low-resolution image representations [13].”
BackgroundA Survey on Visual Place Recognition for Mobile Robots Localization
Chen, Y., Gan, W., Zhang, L., Liu, C., & Wang, X. (2017) · Web Information Systems and Applications Conference (WISA) · doi:10.1109/wisa.2017.7 · Cited by 11
“In recent years, researchers test the utility of visual place recognition by using convolutional neural networks (ConvNet) and sequential image search, which are represented by N Sünderhauf [17] and M Milford [18].”
BackgroundMethodInfluentialJoining the dots: reconstructing 3D environments and movement paths using animal-borne devices
McClune, D. W. (2018) · Animal Biotelemetry · doi:10.1186/s40317-018-0150-6 · Cited by 10
“The application of such techniques in outdoor environments, across seasons, under a range of weather conditions at different times of day is challenging and subject to active research [75, 76].”
Zeng, F., Jacobson, A., Smith, D., Boswell, N., Peynot, T., & Milford, M. J. (2019) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2019.2932579 · Cited by 10
Self-citationZheng, Z., Yang, J., Yu, Z., Wang, Y., Sun, Z., & Zheng, B. (2022) · Neural Networks · doi:10.1016/j.neunet.2022.01.013 · Cited by 12
Why ORB-SLAM is missing commonly occurring loop closures?
Khaliq, S., Anjum, M. L., Hussain, W., Khattak, M. U., & Rasool, M. (2023) · Autonomous Robots · doi:10.1007/s10514-023-10149-x · Cited by 10
“Consequently,amazingvPRsystems have been developed which can work in extreme conditions (day-night (Anoosheh et al., 2019; Milford & Wyeth, 2012), virtual-real world (Zhang et al., 2019), across seasons (Kenshimov et al., 2017; Naseer et al., 2015) and dynamic environments (Merrill & Huang, 2018)).”
BackgroundSurvey Registration for Long‐Term Natural Environment Monitoring
Griffith, S., & Pradalier, C. (2016) · Journal of Field Robotics · doi:10.1002/rob.21664 · Cited by 12
“These factors contribute to its poor performance on our dataset (using the OpenSeqSlam implementation [Milford and Wyeth, 2012]).”
MethodBruce, J., Jacobson, A., & Milford, M. J. (2017) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2017.2724146 · Cited by 10
“In algorithms such as SeqSLAM [2] and SMART [3], a fixed-length window of recent sensory data is matched against stored observations in order to localize a mobile agent.”
BackgroundMethodInfluentialSelf-citationReal-time Visual-Based Localization for Mobile Robot Using Structured-View Deep Learning
Hong, Y.-F., Chang, Y.-M., & Li, C.-H. G. (2019) · IEEE International Conference on Automation Science and Engineering (CASE) · doi:10.1109/coase.2019.8842974 · Cited by 10
An End-to-End Trainable Multi-Column CNN for Scene Recognition in Extremely Changing Environment
Li, Z., Zhou, A., & Shen, Y. (2020) · Sensors · doi:10.3390/s20061556 · Cited by 10
“The orange curve sho s atching based on etho of SeqS as escribe in [31].”
MethodMaltar, J., Marković, I., & Petrovic, I. K. (2020) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2020.103598 · Cited by 12
“SeqSLAM [32] and the proposed approach rely heavily on this fact.”
MethodInfluentialProceedings of the 1st Workshop on Robotics Challenges and Vision (RCV2013)Preprint
Aladren, A., Bodiroža, S., Chitsaz, H., Guerrero, J. J., Hafner, V. V., Hauser, K. K., Jevtić, A., Kazemi, M., Lara, B., Nicolás, G. L., et al. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1402.3213 · Cited by 11
“Other areas in the field of Artificial Intelligence facing NP-hard problems have embraced this paradigm, such as image labeling and machine translation - why not social co-robots?”
BackgroundMethodInfluentialMVP: Unified Motion and Visual Self-Supervised Learning for Large-Scale Robotic NavigationPreprint
Chancán, M., & Milford, M. J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.00667 · Cited by 10
“single-frame methods to work on image sequences [53]– [57]; typically found in driving datasets [11], [12], [58]–[63].”
Self-citationRING#: PR-By-PE Global Localization With Roto-Translation Equivariant Gram Learning
Lu, S., Xu, X., Zhang, D., Wu, Y., Lu, H., Chen, X., Xiong, R., & Wang, Y. (2025) · IEEE Transactions on Robotics · doi:10.1109/tro.2025.3543267 · Cited by 14
“PR [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18] retrieves the map keyframe most similar to the query 1941-0468 © 2025 IEEE.”
Zou, X., Li, J., Wu, W., Liang, F., Yang, B., & Dong, Z. (2025) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2025.03.029 · Cited by 9
Robust Vision-Based Autonomous Navigation, Mapping and Landing for MAVs at Night
Daftry, S., Das, M., Delaune, J. H., Sorice, C., Hewitt, R. A., Reddy, S., Lytle, D., Gu, E., & Matthies, L. H. (2020) · Springer proceedings in advanced robotics · doi:10.1007/978-3-030-33950-0_21 · Cited by 9
“Other approaches [20,18] have used consumer cameras at night, but rely on strong priors from the urban environment.”
BackgroundAppearance-based Loop Closure Detection via Bidirectional Manifold Representation Consensus
Zhang, K., Li, Z., & Ma, J. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9561704 · Cited by 9
“Table III shows the ultimate re-sults of the proposed algorithm and comparative algorithms as reported in the corresponding articles, including SeqS-LAM [30], Bampis et al. [31], PREVIeW [32] and Tsintotas et al. [14].”
The Revisiting Problem in Simultaneous Localization and Mapping
Tsintotas, K. A., Bampis, L., & Gasteratos, A. C. (2022) · Springer tracts in advanced robotics · doi:10.1007/978-3-031-09396-8_1 · Cited by 12
Automatic Coverage Selection for Surface-Based Visual Localization
Mount, J., Dawes, L., & Milford, M. (2019) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2019.2928259 · Cited by 7
“This has been shown to improve visual localization performance [18].”
BackgroundMethodInfluentialSelf-citationLi, Z., Zhou, A., Wang, M., & Shen, Y. (2019) · IEEE International Conference on Robotics and Biomimetics (ROBIO) · doi:10.1109/robio49542.2019.8961602 · Cited by 9
“However, compared with the above methods, AlexNet and SeqSLAM perform poorly.”
MethodInfluentialAutomatic vocabulary and graph verification for accurate loop closure detection
Yue, H., Miao, J., Chen, W., Wang, W., Guo, F., & Li, Z. G. (2022) · Journal of Field Robotics · doi:10.1002/rob.22088 · Cited by 10
“State-of-the-art visual solutions introduce temporal consistency [10]–[12], [28], [29] or geometrical consistency [7], [11] constraints to reject incorrect hypotheses.”
BackgroundDSC: Deep Scan Context Descriptor for Large-Scale Place Recognition
Cui, J., Cai, Y., Huang, T., Zhao, J., Xiong, L., & Yu, Z. (2022) · IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI) · doi:10.1109/mfi55806.2022.9913850 · Cited by 9
VEFNet: an Event-RGB Cross Modality Fusion Network for Visual Place Recognition
Huang, Z., Huang, R., Sun, L., Zhao, C., Huang, M., & Su, S. (2022) · Proceedings - International Conference on Image Processing · doi:10.1109/icip46576.2022.9897492 · Cited by 10
“Using the raw event-based data, the method [3] firstly accumulates events under 10ms temporal resolution into a single frame, and then fed them into the SeqSLAM [7] framework to solve the VPR problem.”
MethodHighly-Efficient Binary Neural Networks for Visual Place Recognition
Ferrarini, B., Milford, M. J., McDonald-Maier, K. D., & Ehsan, S. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981978 · Cited by 12
“A pre-trained AlexNet model [17] on Imagenet [18] is used for loop-closure detection [19], [10] and for enhancing the viewpoint tolerance of SeqSLAM [20], [21].”
Self-citationNeighborhood Manifold Preserving Matching for Visual Place Recognition
Ye, X., & Ma, J. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3216619 · Cited by 9
“SeqSLAM [17] is a classic sequence-based method, which computes the best candidate matching location for every local navigation sequence to achieve the condition-invariant matching, leading to robustness to extreme weather variations.”
MethodDeep causal learning for robotic intelligence
Li, Y. (2023) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2023.1128591 · Cited by 11
Fully Unsupervised Domain-Agnostic Image Retrieval
Zheng, Z., Ren, H., Wu, Y., Zhang, W., Lu, H., Yang, Y., & Shen, H. T. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3335147 · Cited by 17
“This task involves matching query and gallery images from different domains, such as those affected by variations in illumination [18], weather conditions [19], or artistic style [20].”
BackgroundA chordiogram image descriptor using local edgels
Wang, X., Zhang, H., & Peng, G. (2017) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2017.09.005 · Cited by 9
Learning of Holism-Landmark Graph Embedding for Place Recognition in Long-Term Autonomy
Han, F., Elbeleidy, S., Wang, H., Ye, C., & Zhang, H. (2018) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2018.2856274 · Cited by 9
“Similar to image-based scene matching, pairwise sequence similarity scoring is also applied for the place recognition [20], [21], [26].”
BackgroundMethodInfluential
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.