Visual place recognition with CNNs: From global to partial

Xin, Zhe; Cui, Xiaoguang; Zhang, Jixiang; Yang, Yiping; Wang, Yanqing

doi:10.1109/ipta.2017.8310121

Cited by 7 publications

(12 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An initial set of reference image candidates is obtained based on nearest neighbor distances of imagewise global descriptors (Camara et al 2020;B. Liu et al 2021;Xin et al 2017), while local features are used for obtaining a more accurate estimation based on spatial matching (Camara et al 2020;Xin et al 2017) Given that feature maps can extract different types of features depending on the deepness of the respective layers, J. Zhu et al 2018 extracts features from three layers (conv3-3, conv4-4, conv5-3 ) of a VGG16 network and concatenates these to form a global descriptor for an image.…”

Section: Convolutional Neural Network (Cnn)mentioning

confidence: 99%

“…Furthermore, the random selection of descriptor components to reduce its dimensions is employed in Naseer, Oliveira, et al 2017 andXin et al 2017 due to not requiring a learning phase, unlike PCA. Naseer, Oliveira, et al 2017 uses Sparse Random Projection for embedding its FastNet-based descriptor with approximately 130000 dimensions into a reduced 4096-dim descriptor.…”

Section: Descriptor Dimension Reductionmentioning

confidence: 99%

“…The compression procedure achieved similar f-score, and precision-recall metrics compared to the uncompressed descriptor. Xin et al 2017 evaluated the compressing of the 64896-dim AlexNet-based feature used in their work using a random selection of the descriptor's components. The results showed similar precision while speeding up 17 times the matching speed and achieving a compression ratio of 93.7% considering descriptors with 4096 dimensions.…”

Section: Descriptor Dimension Reductionmentioning

confidence: 99%

See 2 more Smart Citations

A Systematic Literature Review on Long-Term Localization and Mapping for Mobile Robots

Sousa

Sobreira

Moreira

2022

Preprint

View full text Add to dashboard Cite

Long-term operation of autonomous robots creates new challenges to the Simultaneous Localization and Mapping (SLAM). Varying conditions of the vehicle’s surroundings, such as appearance variations (lighting, daytime, weather, or seasonal) or reconfigurations of the environment, are a challenge for SLAM algorithms to adapt to new changes while preserving old states. When also operating for long periods and trajectory lengths, the map should readjust to environment changes but not grow indefinitely, where the map size should be dependent only on the explored environment area. Long-term SLAM intends to overcome the challenges associated with lifelong autonomy and improve the robustness of autonomous systems. Although several studies review SLAM algorithms, none of them focus on lifelong autonomy. Thus, this paper presents a systematic literature review on long-term localization and mapping following the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) guidelines. The review analyzes 142 works covering appearance invariance, modeling the environment dynamics, map size management, multi-session, and computational issues including parallel computing and timming efficiency. The analysis also focus on the experimental data and evaluation metrics commonly used to assess long-term autonomy. Moreover, an overview over the bibliographic data of the 142 records provides analysis in terms of keywords and authorship co-occurrence to identify the terms more used in long-term SLAM and research networks between authors, respectively. Future studies can update this paper thanks to the systematic methodology presented in the review and the public GitHub repository with all the documentation and scripts used during the review process.

show abstract

Section: Convolutional Neural Network (Cnn)mentioning

confidence: 99%

Section: Descriptor Dimension Reductionmentioning

confidence: 99%

Section: Descriptor Dimension Reductionmentioning

confidence: 99%

See 1 more Smart Citation

A Systematic Literature Review on Long-Term Localization and Mapping for Mobile Robots

Sousa

Sobreira

Moreira

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…The task of coarsely estimating the place where a photo was taken based on a set of previously visited locations is called Visual (Image) Geo-localization (VG) [37,42,86] or Visual Place Recognition (VPR) [21,44] and it is addressed using image matching and retrieval methods on a database of images of known locations. We are witnessing a rapid growth of this field of research, as demonstrated by the increasing number of publications [2,10,14,[22][23][24]28,30,36,37,42,44,46,57,60,72,73,[76][77][78][79]81,87], but this expansion is accompanied by two major limitations: i) A focus on single metric optimization, as it is common practice to compare results solely based on the recall on chosen datasets and ignoring other factors such as execution time, hardware requirements, and scalability. All these aspects are important constraints in the design of a real-world VG system.…”

Section: Introductionmentioning

confidence: 99%

Deep Visual Geo-localization Benchmark

Gabriele¹,

Mereu²,

Gabriele³

et al. 2022

Preprint

View full text Add to dashboard Cite

In this paper, we propose a new open-source benchmarking framework for Visual Geo-localization (VG) that allows to build, train, and test a wide range of commonly used architectures, with the flexibility to change individual components of a geo-localization pipeline. The purpose of this framework is twofold: i) gaining insights into how different components and design choices in a VG pipeline impact the final results, both in terms of performance (recall@N metric) and system requirements (such as execution time and memory consumption); ii) establish a systematic evaluation protocol for comparing different methods. Using the proposed framework, we perform a large suite of experiments which provide criteria for choosing backbone, aggregation and negative mining depending on the use-case and requirements. We also assess the impact of engineering techniques like pre/post-processing, data augmentation and image resizing, showing that better performance can be obtained through somewhat simple procedures: for example, downscaling the images' resolution to 80% can lead to similar results with a 36% savings in extraction time and dataset storage requirement. Code and trained models are available at https://deep-vg-bench.herokuapp.com/.

show abstract

“…It can generally be extended to broader areas, including topological mapping, loop closure and drift removal in geometric mapping and learning scene dynamics for long-term localization and mapping. Long-term operations in environments can cause significant image variations including illumination changes, occlusion and scene dynamics [2].…”

Section: Introductionmentioning

confidence: 99%

Supervised Fine-tuning Evaluation for Long-term Visual Place Recognition

Alijani

Rahtu

2021

2021 IEEE 23rd International Workshop on Multimedia Signal Processing (MMSP)

View full text Add to dashboard Cite

In this paper, we present a comprehensive study on the utility of deep convolutional neural networks with two stateof-the-art pooling layers which are placed after convolutional layers and fine-tuned in an end-to-end manner for visual place recognition task in challenging conditions, including seasonal and illumination variations. We compared extensively the performance of deep learned global features with three different loss functions, e.g. triplet, contrastive and ArcFace, for learning the parameters of the architectures in terms of fraction of the correct matches during deployment. To verify effectiveness of our results, we utilized two real world datasets in place recognition, both indoor and outdoor. Our investigation demonstrates that fine tuning architectures with ArcFace loss in an end-to-end manner outperforms other two losses by approximately 1 ∼ 4 % in outdoor and 1 ∼ 2 % in indoor datasets, given certain thresholds, for the visual place recognition tasks.

show abstract

Visual place recognition with CNNs: From global to partial

Cited by 7 publications

References 14 publications

A Systematic Literature Review on Long-Term Localization and Mapping for Mobile Robots

A Systematic Literature Review on Long-Term Localization and Mapping for Mobile Robots

Deep Visual Geo-localization Benchmark

Supervised Fine-tuning Evaluation for Long-term Visual Place Recognition

Contact Info

Product

Resources

About