Two-stage Discriminative Re-ranking for Large-scale Landmark Retrieval

Yokoo, Shuhei; Ozaki, Kohei; Simo-Serra, Edgar; Iizuka, Satoshi

doi:10.1109/cvprw50498.2020.00514

“…III-A). For instance, [67], [88] use the ArcFace loss [89] to train the global features with image level labels, achieving good retrieval results under the cosine similarity. There is a couple of reason that motivate this interest in revisiting classification models as generators of image representations for VPR.…”

Section: A Learning From Classificationmentioning

confidence: 99%

“…The shortlisted images are re-ranked based on the maximum similarity between their regions and the query. In [88], [136] the authors use a discriminative ranking method based on the similarity of labels assigned to the images by a kNN search with soft voting. Namely, the search results are re-ranked by first moving up all the shortlisted images that have the same label as the query, and then by adding the images from the database with the same label as the query and that were not retrieved by the search.…”

Section: B Non-geometric Re-rankingmentioning

confidence: 99%

“…One of the most successful and widely used techniques to improve the retrieval result is query expansion (QE) [12], [23], [37], [65], [66], [70], [73], [74], [76], [77], [80], [82], [88], [92], [96], [123], [126], [127], [129], [130], [138]. The idea of query expansion is to use the shortlisted images as a feedback to produce an enriched representation that is re-submitted for a new search through the database.…”

Section: Query Expansionmentioning

confidence: 99%

See 1 more Smart Citation

A Survey on Deep Visual Place Recognition

Masone

¹

,

Caputo

²

2021

View full text Add to dashboard Cite

In recent years visual place recognition (VPR), i.e., the problem of recognizing the location of images, has received considerable attention from multiple research communities, spanning from computer vision to robotics and even machine learning. This interest is fueled on one hand by the relevance that visual place recognition holds for many applications and on the other hand by the unsolved challenge of making these methods perform reliably in different conditions and environments. This paper presents a survey of the state-of-the-art of research on visual place recognition, focusing on how it has been shaped by the recent advances in deep learning. We start discussing the image representations used in this task and how they have evolved from using hand-crafted to deep-learned features. We further review how metric learning techniques are used to get more discriminative representations, as well as techniques for dealing with occlusions, distractors, and shifts in the visual domain of the images. The survey also provides an overview of the specific solutions that have been proposed for applications in robotics and with aerial imagery. Finally the survey provides a summary of datasets that are used in visual place recognition, highlighting their different characteristics. INDEX TERMS Visual place recognition, image representation learning, deep learning.

show abstract

“…The works in the first category focus on the localization on fine granularity scales, such as landmarks [3,5,46,49] or at city-scale granularity [1,17,25,39,44]. In general, the solutions that are employed for such problems are based on retrieval systems that match the query images with ones from a background collection and then apply a post-processing scheme to estimate the final location.…”

Section: Related Workmentioning

confidence: 99%

Leveraging EfficientNet and Contrastive Learning for Accurate Global-scale Location Estimation

Kordopatis-Zilos

¹

,

Galopoulos

²

,

Papadopoulos

³

et al. 2021

Proceedings of the 2021 International Conference on Multimedia Retrieval

View full text Add to dashboard Cite

In this paper, we address the problem of global-scale image geolocation, proposing a mixed classification-retrieval scheme. Unlike other methods that strictly tackle the problem as a classification or retrieval task, we combine the two practices in a unified solution leveraging the advantages of each approach with two different modules. The first leverages the EfficientNet architecture to assign images to a specific geographic cell in a robust way. The second introduces a new residual architecture that is trained with contrastive learning to map input images to an embedding space that minimizes the pairwise geodesic distance of same-location images. For the final location estimation, the two modules are combined with a search-within-cell scheme, where the locations of most similar images from the predicted geographic cell are aggregated based on a spatial clustering scheme. Our approach demonstrates very competitive performance on four public datasets, achieving new state-of-the-art performance in fine granularity scales, i.e., 15.0% at 1km range on Im2GPS3k. CCS CONCEPTS• Computing methodologies → Computer vision problems;• Information systems → Geographic information systems.

show abstract

“…The works in the first category focus on the localization on fine granularity scales, such as landmarks [3,5,46,49] or at cityscale granularity [1,17,25,39,44]. In general, the solutions that are employed for such problems are based on retrieval systems that match the query images with ones from a background collection and then apply a post-processing scheme to estimate the final location.…”

Section: Related Workmentioning

confidence: 99%

Leveraging EfficientNet and Contrastive Learning for Accurate Global-scale Location Estimation

Kordopatis-Zilos¹,

Galopoulos²,

Papadopoulos³

et al. 2021

Preprint

0

View full text Add to dashboard Cite

In this paper, we address the problem of global-scale image geolocation, proposing a mixed classification-retrieval scheme. Unlike other methods that strictly tackle the problem as a classification or retrieval task, we combine the two practices in a unified solution leveraging the advantages of each approach with two different modules. The first leverages the EfficientNet architecture to assign images to a specific geographic cell in a robust way. The second introduces a new residual architecture that is trained with contrastive learning to map input images to an embedding space that minimizes the pairwise geodesic distance of same-location images. For the final location estimation, the two modules are combined with a search-within-cell scheme, where the locations of most similar images from the predicted geographic cell are aggregated based on a spatial clustering scheme. Our approach demonstrates very competitive performance on four public datasets, achieving new state-of-the-art performance in fine granularity scales, i.e., 15.0% at 1km range on Im2GPS3k.

show abstract

Two-stage Discriminative Re-ranking for Large-scale Landmark Retrieval

Cited by 24 publications

References 40 publications

A Survey on Deep Visual Place Recognition

A Survey on Deep Visual Place Recognition

Leveraging EfficientNet and Contrastive Learning for Accurate Global-scale Location Estimation

Leveraging EfficientNet and Contrastive Learning for Accurate Global-scale Location Estimation

Contact Info

Product

Resources

About