369 Hits in 5.6 sec

Large-Scale Location Recognition and the Geometric Burstiness Problem

Torsten Sattler, Michal Havlena, Konrad Schindler, Marc Pollefeys
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Location recognition is often cast as an image retrieval problem and recent research has almost exclusively focused on improving the chance that a relevant database image is ranked high enough after retrieval  ...  Our approach is easy to implement and can easily be integrated into existing location recognition systems.  ...  This work was supported by Google's Project Tango and EC Horizon 2020 project REPLICATE (no. 687757). The authors thank Relja Arandjelović for his invaluable help with the DisLoc algorithm.  ... 
doi:10.1109/cvpr.2016.175 dblp:conf/cvpr/SattlerHSP16 fatcat:gke7v6j3orh3vp42huvytkwshq

Visual Place Recognition with Repetitive Structures

Akihiko Torii, Josef Sivic, Masatoshi Okutomi, Tomas Pajdla
2015 IEEE Transactions on Pattern Analysis and Machine Intelligence  
baseline as well as the more recently proposed burstiness weighting and Fisher vector encoding.  ...  Place recognition results are shown on datasets of street-level imagery from Pittsburgh and San Francisco demonstrating significant gains in recognition performance compared to the standard bag-of-visual-words  ...  and the ERC project LEAP.  ... 
doi:10.1109/tpami.2015.2409868 pmid:26440272 fatcat:kqh7wpcc2be4hb6wsej2yy57we

Grassmann Pooling as Compact Homogeneous Bilinear Pooling for Fine-Grained Visual Classification [chapter]

Xing Wei, Yue Zhang, Yihong Gong, Jiawei Zhang, Nanning Zheng
2018 Lecture Notes in Computer Science  
Designing discriminative and invariant features is the key to visual recognition.  ...  However, the bilinear feature could suffer from the visual burstiness phenomenon similar to other visual representations such as VLAD and Fisher Vector.  ...  Introduction Visual recognition problems mainly have two challenges: between-class similarity and within-class variance.  ... 
doi:10.1007/978-3-030-01219-9_22 fatcat:3ipzmurq7fcvvgzwcee3ddpf6e

Local Feature Reliability Measure Consistent with Match Conditions for Mobile Visual Search

2018 IEICE transactions on information and systems  
These features are selected according to our novel reliability measure which enables robust recognition against various viewpoint changes. We regard feature selection as a maximum coverage problem.  ...  As this problem is NP-hard and thus computationally infeasible, we explore approximate solutions based on a greedy algorithm.  ...  This is a modified version of the local NBNN scoring scheme [26] . Weak geometric consistency. Weak geometric consistency (WGC) provides constraints based on angle and scale information [21] .  ... 
doi:10.1587/transinf.2018edp7107 fatcat:ds66cuo3crfmnancsc53nage4i

Lp-Norm IDF for Large Scale Image Search

Liang Zheng, Shengjin Wang, Ziqiong Liu, Qi Tian
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition  
We report a performance improvement of as large as 27.1% over the baseline approach.  ...  To tackle this problem, this paper introduces a novel IDF expression by the use of L p -norm pooling technique.  ...  We couple both the inter-and intra-image burstiness solutions in our experiment. We select the best-performance formula in [5] .  ... 
doi:10.1109/cvpr.2013.213 dblp:conf/cvpr/ZhengWLT13 fatcat:nuvjovewyngmhje5b53rezgi3i

Visual Place Recognition with Repetitive Structures

Akihiko Torii, Josef Sivic, Toma Pajdla, Masatoshi Okutomi
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition  
baseline and more recently proposed burstiness weighting.  ...  Place recognition results are shown on datasets of street-level imagery from Pittsburgh and San Francisco demonstrating significant gains in recognition performance compared to the standard bag-of-visual-words  ...  In addition, the proposed method is simple and can be easily incorporated into existing large scale place recognition architectures.  ... 
doi:10.1109/cvpr.2013.119 dblp:conf/cvpr/ToriiSPO13 fatcat:rxq6nhg55rgy5h24sct2t3scru

Adding spatial distribution clue to aggregated vector in image retrieval

Pingping Liu, Zhuang Miao, Huili Guo, Yeran Wang, Ni Ai
2018 EURASIP Journal on Image and Video Processing  
For each cluster, the corresponding spatial distribution is computed using a histogram of spatial locations, scales, and orientations of all local features inside the cluster.  ...  Entropy is calculated from the spatial distributions of all clusters of an image to create a distribution function, which is further normalized and concatenated with the VLAD vector to generate the final  ...  ZM finished the combination of our method with VLAD and completed the revision of the paper. HG, YW, and NA validated the results. All authors read and approved the final manuscript.  ... 
doi:10.1186/s13640-018-0247-0 fatcat:qnhllkvjrncx3o7fyqd6nuy2y4

Pairwise geometric matching for large-scale object retrieval

Xinchao Li, Martha Larson, Alan Hanjalic
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
for large-scale object retrieval.  ...  In addition, we combine the information on geometric relations from both the individual correspondences and pairs of correspondences to further improve the verification accuracy.  ...  Avrithis and Tolias [2] exploited the relative geometric relations, i.e., scaling, orientation and location, between the local appearance of the matched points.  ... 
doi:10.1109/cvpr.2015.7299151 dblp:conf/cvpr/LiLH15 fatcat:enjbpf25rzb6xewusogiwxn4mm

DisLocation: Scalable Descriptor Distinctiveness for Location Recognition [chapter]

Relja Arandjelović, Andrew Zisserman
2015 Lecture Notes in Computer Science  
The objective of this paper is to improve large scale visual object retrieval for visual place recognition.  ...  The method is evaluated on standard publicly available large-scale place recognition benchmarks containing street-view imagery of Pittsburgh and San Francisco.  ...  Torri and D. Chen for sharing their datasets, and are grateful for financial support from ERC grant VisRec no. 228180 and a Royal Society Wolfson Research Merit Award.  ... 
doi:10.1007/978-3-319-16817-3_13 fatcat:p7b75vk3kvdndfbzglzskwcpt4

Hyperpoints and Fine Vocabularies for Large-Scale Location Recognition

Torsten Sattler, Michal Havlena, Filip Radenovic, Konrad Schindler, Marc Pollefeys
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
Recently, several authors have tried to overcome these problems, either by a smart compression of the 3D model or by clever sampling strategies for geometric verification.  ...  While this is almost trivial at small scale, special care must be taken as the size of the 3D model grows, because straight-forward descriptor matching becomes ineffective due to the large memory footprint  ...  The advantage of our approach is that it easily scales to large datasets since adding a point and its descriptors requires only a few additional bytes.  ... 
doi:10.1109/iccv.2015.243 dblp:conf/iccv/SattlerHRSP15 fatcat:cxkfrxz3v5baxfjqeh5iclx5im

Are Large-Scale 3D Models Really Necessary for Accurate Visual Localization?

Akihiko Torii, Hajime Taira, Josef Sivic, Marc Pollefeys, Masatoshi Okutomi, Tomas Pajdla, Torsten Sattler
2019 IEEE Transactions on Pattern Analysis and Machine Intelligence  
However, constructing (and extending) large-scale 3D models is still a significant challenge.  ...  We create reference poses for a large and challenging urban dataset.  ...  We use the combination of Disloc with the geometric burstiness weighting scheme recently proposed in [37] .  ... 
doi:10.1109/tpami.2019.2941876 pmid:31535984 fatcat:y7ii3bxoezd7rh5k5jz7t5arla

Robust Visual Place Recognition with Graph Kernels

Elena Stumm, Christopher Mei, Simon Lacroix, Juan Nieto, Marco Hutter, Roland Siegwart
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Furthermore, the place recognition complexity is independent of the size of the map.  ...  A novel method for visual place recognition is introduced and evaluated, demonstrating robustness to perceptual aliasing and observation noise.  ...  Acknowledgments The research leading to these results has received funding from the Swiss National Science Foundation through the National Centre of Competence in Research Robotics.  ... 
doi:10.1109/cvpr.2016.491 dblp:conf/cvpr/StummMLNHS16 fatcat:js4dnny4rzampgdcg5j6puufuq

Local Feature Detectors, Descriptors, and Image Representations: A Survey [article]

Yusuke Uchida
2016 arXiv   pre-print
With the advances in both stable interest region detectors and robust and distinctive descriptors, local feature-based image or object retrieval has become a popular research topic.  ...  %All of the local feature-based image retrieval system involves two important processes: local feature extraction and image representation.  ...  Introduction Image retrieval is the problem of searching for digital images in large databases.  ... 
arXiv:1607.08368v1 fatcat:g3qm64yonvcd5fe5sw4zbn3mo4

Person Re-identification Meets Image Search [article]

Liang Zheng, Liyue Shen, Lu Tian, Shengjin Wang, Jiahao Bu, Qi Tian
2015 arXiv   pre-print
In the light of recent advances in image search, this paper proposes to treat person re-identification as an image search problem.  ...  However, for person re-identification, the effectiveness of local features and the "query-search" mode make it well posed for image search techniques.  ...  Although good recognition rate is achieved, this line of methods suffer from low computational efficiency, which limits its potential in large-scale applications.  ... 
arXiv:1502.02171v1 fatcat:l6ii2i3uyvftfjz6yvzvub7e4q

Distribution Entropy Boosted VLAD for Image Retrieval

Qiuzhan Zhou, Cheng Wang, Pingping Liu, Qingliang Li, Yeran Wang, Shuozhang Chen
2016 Entropy  
The original VLAD adopts only residuals to depict the distribution information of every visual word and neglects other statistical clues, so its discriminative power is limited.  ...  We validate our approach in image retrieval and image classification experiments. Experimental results demonstrate the effectiveness of our algorithm.  ...  Cheng Wang, Qingliang Li, Yeran Wang and Shuozhang Chen validated the results. All authors have read and approved the final manuscript.  ... 
doi:10.3390/e18080311 fatcat:sih3ua4zfzhmtocfr3uipcob4u
« Previous Showing results 1 — 15 out of 369 results