Restoring Super-High Resolution GPS Mobility Data

Hamada Rizk; Haruki Yonekura; Hirozumi Yamaguchi; Ren Ozeki

arxiv: 2410.12818 · v1 · submitted 2024-10-01 · 📡 eess.SP · cs.LG

Restoring Super-High Resolution GPS Mobility Data

Haruki Yonekura , Ren Ozeki , Hamada Rizk , Hirozumi Yamaguchi This is my paper

Pith reviewed 2026-05-23 20:13 UTC · model grok-4.3

classification 📡 eess.SP cs.LG

keywords GPS trajectory reconstructiontransformer encoder-decodergraph convolutional networksprivacy-preserving mobility dataFréchet distance evaluationBeijing trajectorieshigh-resolution restorationtruncated GPS data

0 comments

The pith

A transformer-GCN model reconstructs high-resolution GPS trajectories from truncated low-resolution inputs with 0.198 km average Fréchet error.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents a system that combines transformer encoder-decoder models with graph convolutional networks to restore fine-grained GPS mobility trajectories from low-resolution versions created by truncation or rounding. This targets the tension between protecting user privacy through coarsened data and retaining enough detail for useful mobility analysis. Evaluation on the Beijing trajectory dataset shows the model recovers trajectories with lower error than map-matching or LSTM-based synthetic generation approaches. A reader would care because the approach could let applications keep working with detailed path information even after privacy steps are applied. The system also handles both real and synthetic inputs without major loss of performance.

Core claim

The proposed system integrates transformer-based encoder-decoder models with graph convolutional networks to capture both the temporal dependencies of trajectory data and the spatial relationships in road networks. By doing so it recovers fine-grained trajectory details lost through common truncation or rounding practices used for privacy. On the Beijing trajectory dataset the model reaches an average Fréchet distance of 0.198 km, compared with 0.632 km for map-matching algorithms and 0.498 km for synthetic trajectory models, and it generalizes to synthetic data as well.

What carries the argument

Transformer-based encoder-decoder paired with graph convolutional networks that jointly process temporal sequence patterns and road-network spatial structure to reconstruct trajectories.

If this is right

The system can be deployed in urban mobility applications while maintaining both reconstruction accuracy and privacy protection.
It recovers real-world trajectories more accurately than map-matching or LSTM synthetic methods.
Performance holds when the input is itself synthetic rather than real truncated data.
The hybrid architecture exploits both sequence modeling and graph structure to close the resolution gap created by privacy measures.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same architecture could be adapted to restore other coarsened location streams such as check-in data or ride-hailing traces.
Running the restoration on-device rather than centrally might further reduce privacy exposure by avoiding transmission of even the low-resolution input.
Performance on road networks with very different topologies would test whether the GCN component generalizes beyond the training city.
Combining the model with differential privacy mechanisms at the truncation stage could produce a tunable privacy-utility curve.

Load-bearing premise

The Beijing trajectory dataset together with the chosen truncation and rounding steps stand in for the range of privacy-protected mobility data encountered in practice.

What would settle it

Running the trained model on a trajectory collection from another city that uses different truncation thresholds and measuring whether the average Fréchet distance stays near 0.2 km.

Figures

Figures reproduced from arXiv: 2410.12818 by Hamada Rizk, Haruki Yonekura, Hirozumi Yamaguchi, Ren Ozeki.

**Figure 1.** Figure 1: System overview. 3 SYSTEM DETAILS The system is designed to reconstruct high-resolution GPS trajectory data from low-resolution, privacy-preserved inputs. As shown in [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

**Figure 2.** Figure 2: Data publication scenarios. 4 EVALUATION The evaluation of the proposed system was conducted to assess its ability to reconstruct high-resolution GPS trajectories from lowresolution. This section provides a detailed explanation of the experimental settings, evaluation metrics, and the results obtained. The experiments focus on the performance comparison between the proposed model and traditional methods… view at source ↗

**Figure 3.** Figure 3: Comparison of reconstructed trajectories, error histograms, and visualizations. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

read the original abstract

This paper presents a novel system for reconstructing high-resolution GPS trajectory data from truncated or synthetic low-resolution inputs, addressing the critical challenge of balancing data utility with privacy preservation in mobility applications. The system integrates transformer-based encoder-decoder models with graph convolutional networks (GCNs) to effectively capture both the temporal dependencies of trajectory data and the spatial relationships in road networks. By combining these techniques, the system is able to recover fine-grained trajectory details that are lost through data truncation or rounding, a common practice to protect user privacy. We evaluate the system on the Beijing trajectory dataset, demonstrating its superior performance over traditional map-matching algorithms and LSTM-based synthetic data generation methods. The proposed model achieves an average Fr\'echet distance of 0.198 km, significantly outperforming map-matching algorithms (0.632 km) and synthetic trajectory models (0.498 km). The results show that the system is not only capable of accurately reconstructing real-world trajectories but also generalizes effectively to synthetic data. These findings suggest that the system can be deployed in urban mobility applications, providing both high accuracy and robust privacy protection.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The transformer-GCN model reports lower Fréchet distance on restored Beijing trajectories than the baselines, but the gains rest on point estimates with no variance or tests shown.

read the letter

The main takeaway is that their hybrid model reaches 0.198 km average Fréchet distance on restored GPS paths from truncated inputs, beating map-matching at 0.632 km and the synthetic baseline at 0.498 km. That number is the concrete result to check first. The approach itself combines a transformer encoder-decoder with GCN layers to handle sequence dependencies and road-network structure at the same time. This is a straightforward extension of existing sequence and graph models to the specific task of undoing truncation or rounding for privacy reasons. They test on the Beijing dataset for both real truncated traces and synthetic ones, which shows some effort to check generalization within that setting. The integration of temporal and spatial components is described clearly enough in the abstract to see why it might recover finer details than pure map-matching. The soft spots are the missing pieces around the numbers. No standard deviations, no test-set size, no splits, and no significance checks are mentioned, so the gap could shrink or disappear under different partitions or with proper error bars. Everything is also tied to one city and one set of truncation rules, which leaves open how much carries over to other places or privacy schemes. The paper is aimed at people working on mobility data utility under privacy constraints, such as urban planners or location-based service developers who need to reconstruct traces. A reader in that area could pick up the architecture and try it on their own data. It is worth sending to peer review because the task is well-defined, the comparison is direct, and the method is reproducible in principle once the training details are filled in. The statistical reporting will need tightening, but the core idea is solid enough to merit referee time.

Referee Report

2 major / 1 minor

Summary. The paper claims to present a novel system integrating transformer-based encoder-decoder models with graph convolutional networks (GCNs) to reconstruct high-resolution GPS trajectories from truncated or rounded low-resolution inputs for privacy preservation. Evaluated on the Beijing trajectory dataset, it reports an average Fréchet distance of 0.198 km, significantly outperforming map-matching algorithms (0.632 km) and synthetic trajectory models (0.498 km), with effective generalization to synthetic data.

Significance. If the quantitative superiority claims hold under rigorous statistical validation and reproducible experimental protocols, the hybrid architecture could meaningfully advance privacy-utility tradeoffs in mobility data applications by recovering fine-grained spatial-temporal details from coarsened inputs.

major comments (2)

[Abstract and Results] Abstract and Results: The headline performance numbers (average Fréchet distance of 0.198 km vs. 0.632 km and 0.498 km) are presented solely as point estimates with no standard deviations, confidence intervals, sample sizes, bootstrap statistics, or hypothesis tests referenced, leaving open whether the reported gaps exceed variability across trajectories or data partitions.
[Methods] Methods: The manuscript provides no details on training procedures, validation splits, hyperparameter selection, data preprocessing (including exact truncation/rounding mechanics), or model architecture specifics, which are required to assess whether the reported metrics reliably support the central superiority claim.

minor comments (1)

[Abstract] Abstract: The description of the Beijing dataset and the truncation/rounding procedures could be expanded to clarify the evaluation setting and support claims of generalization.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive comments. We address each major point below and commit to revisions that strengthen the statistical rigor and reproducibility of the work.

read point-by-point responses

Referee: [Abstract and Results] Abstract and Results: The headline performance numbers (average Fréchet distance of 0.198 km vs. 0.632 km and 0.498 km) are presented solely as point estimates with no standard deviations, confidence intervals, sample sizes, bootstrap statistics, or hypothesis tests referenced, leaving open whether the reported gaps exceed variability across trajectories or data partitions.

Authors: We agree that the current presentation of results as single point estimates limits the ability to assess variability. In the revised manuscript we will add standard deviations computed across trajectory partitions or multiple random seeds, 95% confidence intervals, the exact number of test trajectories, and the results of paired statistical tests (e.g., Wilcoxon signed-rank) comparing our model against the baselines. These additions will be placed in both the abstract and the Results section. revision: yes
Referee: [Methods] Methods: The manuscript provides no details on training procedures, validation splits, hyperparameter selection, data preprocessing (including exact truncation/rounding mechanics), or model architecture specifics, which are required to assess whether the reported metrics reliably support the central superiority claim.

Authors: We acknowledge that the Methods section is currently insufficient for reproducibility. We will expand it to describe: the optimizer and learning-rate schedule, number of epochs and early-stopping criterion, the train/validation/test split ratios and any stratification, the hyperparameter search procedure, the precise truncation and rounding operations applied to latitude/longitude (including the rounding granularity in meters), and the full architectural specifications (number of transformer layers, attention heads, GCN hidden dimensions, and fusion mechanism). A new subsection on experimental protocol will be added. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical ML evaluation with no derivations or self-referential fits

full rationale

The paper presents a transformer+GCN model for trajectory restoration and reports empirical Fréchet distances on the Beijing dataset. No equations, derivations, or parameter-fitting steps are described that reduce a claimed result to its own inputs by construction. The performance numbers are direct outputs of model evaluation rather than renamed fits or self-citation chains. The reader's assessment of score 2.0 is consistent with the absence of any load-bearing circular steps; the work is self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no free parameters, axioms, or invented entities are identifiable from the provided text.

pith-pipeline@v0.9.0 · 5727 in / 1061 out tokens · 28778 ms · 2026-05-23T20:13:18.497629+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

30 extracted references · 30 canonical work pages

[1]

Osman Abul, Francesco Bonchi, and Mirco Nanni. 2008. Never Walk Alone: Uncertainty for Anonymity in Moving Objects Databases. In 2008 IEEE 24th International Conference on Data Engineering . 376–385

work page 2008
[2]

Chen Chen, Cewu Lu, Qixing Huang, Qiang Yang, Dimitrios Gunopulos, and Leonidas Guibas. 2016. City-Scale Map Creation and Updating using GPS Collec- tions. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowl- edge Discovery and Data Mining (KDD ’16). Association for Computing Machinery, New York, NY, USA, 1465–1474. https://doi.org/10....

work page doi:10.1145/2939672.2939833 2016
[3]

Yumeki Goto, Tomoya Matsumoto, Hamada Rizk, Naoto Yanai, and Hirozumi Yamaguchi. 2023. Privacy-Preserving Taxi-Demand Prediction Using Federated Learning. In International Conference on Smart Computing . IEEE

work page 2023
[4]

Hemkumar, S

D. Hemkumar, S. Ravichandra, and D.V.L.N. Somayajulu. 2020. Impact of prior knowledge on privacy leakage in trajectory data publishing. Engineering Science and Technology, an International Journal 23, 6 (2020), 1291–1300

work page 2020
[5]

Hongbo Jiang, Jie Li, Ping Zhao, Fanzi Zeng, Zhu Xiao, and Arun Iyengar. 2021. Location Privacy-Preserving Mechanisms in Location-Based Services: A Com- prehensive Survey. ACM Comput. Surv. 54, 1, Article 4 (jan 2021), 36 pages

work page 2021
[6]

Kaifeng Jiang, Dongxu Shao, Stéphane Bressan, Thomas Kister, and Kian-Lee Tan. 2013. Publishing Trajectories with Differential Privacy Guarantees. In Pro- ceedings of the 25th International Conference on Scientific and Statistical Database Management (SSDBM). Association for Computing Machinery, New York, NY, USA, Article 12, 12 pages. https://doi.org/10....

work page doi:10.1145/2484838.2484846 2013
[7]

Yang Li, Yangyan Li, Dimitrios Gunopulos, and Leonidas Guibas. 2016. Knowledge-based trajectory completion from sparse GPS samples. In Proceedings of the 24th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems. 1–10

work page 2016
[8]

Yin Lou, Chengyang Zhang, Yu Zheng, Xing Xie, Wei Wang, and Yan Huang

work page
[9]

In Proceedings of the 17th ACM SIGSPATIAL international conference on advances in geographic information systems

Map-matching for low-sampling-rate GPS trajectories. In Proceedings of the 17th ACM SIGSPATIAL international conference on advances in geographic information systems. 352–361

work page
[10]

Antonios Makris, Camila Leite da Silva, Vania Bogorny, Luis Otavio Alvares, Jose Antonio Macedo, and Konstantinos Tserpes. 2021. Evaluating the effect of compressing algorithms for trajectory similarity and classification problems. Geoinformatica 25, 4 (oct 2021), 679–711

work page 2021
[11]

Olsen, Jeong-Hyon Hwang, Catherine T

Jonathan Muckell, Paul W. Olsen, Jeong-Hyon Hwang, Catherine T. Lawson, and S. S. Ravi. 2014. Compression of trajectory data: a comprehensive evaluation and new approach. Geoinformatica 18, 3 (jul 2014), 435–460

work page 2014
[12]

Mashaal Musleh and Mohamed F Mokbel. 2023. Kamel: A Scalable BERT-based System for Trajectory Imputation. Proceedings of the VLDB Endowment 17, 3 (2023), 525–538

work page 2023
[13]

Paul Newson and John Krumm. 2009. Hidden Markov map matching through noise and sparseness. In Proceedings of the 17th ACM SIGSPATIAL international conference on advances in geographic information systems . 336–343

work page 2009
[14]

Ren Ozeki, Haruki Yonekura, Aidana Baimbetova, Hamada Rizk, and Hirozumi Yamaguchi. 2023. One Model Fits All: Cross-Region Taxi-Demand Forecasting. In Proceedings of the 31st ACM International Conference on Advances in Geographic Information Systems (SIGSPATIAL ’23). Association for Computing Machinery, New York, NY, USA, Article 101, 4 pages. https://doi...

work page doi:10.1145/3589132.3625651 2023
[15]

Ren Ozeki, Haruki Yonekura, Hamada Rizk, and Hirozumi Yamaguchi. 2022. Sharing without caring: privacy protection of users’ spatio-temporal data without compromise on utility. In Proceedings of the 30th International Conference on Advances in Geographic Information Systems . 1–2

work page 2022
[16]

Ren Ozeki, Haruki Yonekura, Hamada Rizk, and Hirozumi Yamaguchi. 2023. Bal- ancing Privacy and Utility of Spatio-Temporal Data for Taxi-Demand Prediction. In The 24th IEEE International Conference on Mobile Data Management . IEEE

work page 2023
[17]

Ren Ozeki, Haruki Yonekura, Hamada Rizk, and Hirozumi Yamaguchi. 2023. Balancing privacy and utility of spatio-temporal data for taxi-demand prediction. In 2023 24th IEEE International Conference on Mobile Data Management (MDM) . IEEE, 215–220

work page 2023
[18]

Ren Ozeki, Haruki Yonekura, Hamada Rizk, and Hirozumi Yamaguchi. 2024. Privacy-Preserved Taxi Demand Prediction System Utilizing Distributed Data. In Proceedings of the 32nd ACM International Conference on Advances in Geographic Information Systems (SIGSPATIAL ’24). Association for Computing Machinery, to appear

work page 2024
[19]

Qin, Yongli Ren, Wei Shao, Brennan Lake, Filippo Privitera, and Flora D

Kyle K. Qin, Yongli Ren, Wei Shao, Brennan Lake, Filippo Privitera, and Flora D. Salim. 2023. Multiple-level Point Embedding for Solving Human Trajectory Imputation with Prediction. ACM Trans. Spatial Algorithms Syst. 9, 2, Article 15 (may 2023), 22 pages. https://doi.org/10.1145/3582427

work page doi:10.1145/3582427 2023
[20]

Jinmeng Rao, Song Gao, Yuhao Kang, and Qunying Huang. 2020. LSTM-TrajGAN: A Deep Learning Approach to Trajectory Privacy Protection. https://doi.org/10. 48550/ARXIV.2006.10521

work page arXiv 2020
[21]

Sijie Ruan, Cheng Long, Jie Bao, Chunyang Li, Zisheng Yu, Ruiyuan Li, Yuxuan Liang, Tianfu He, and Yu Zheng. 2020. Learning to Generate Maps from Trajec- tories. Proceedings of the AAAI Conference on Artificial Intelligence 34, 01 (Apr. 2020), 890–897. https://ojs.aaai.org/index.php/AAAI/article/view/5435

work page 2020
[22]

Akiyoshi Suzuki, Mayu Iwata, Yuki Arase, Takahiro Hara, Xing Xie, and Shojiro Nishio. 2010. A User Location Anonymization Method for Location Based Ser- vices in a Real Environment. In Proceedings of the 18th SIGSPATIAL International Conference on Advances in Geographic Information Systems (GIS ’10) . Association for Computing Machinery, New York, NY, USA...

work page 2010
[23]

Jingyuan Wang, Ning Wu, Xinxi Lu, Wayne Xin Zhao, and Kai Feng. 2021. Deep Trajectory Recovery with Fine-Grained Calibration using Kalman Filter. IEEE Transactions on Knowledge and Data Engineering 33, 3 (2021), 921–934

work page 2021
[24]

Szymon Woźniak and Piotr Szymański. 2021. Hex2vec: Context-aware embedding h3 hexagons with openstreetmap tags. In Proceedings of the 4th ACM SIGSPATIAL International Workshop on AI for Geographic Knowledge Discovery . 61–71

work page 2021
[25]

Haruki Yonekura, Ren Ozeki, Hamada Rizk, and Hirozumi Yamaguchi. 2023. STM-A Privacy-Enhanced Solution for Spatio-Temporal Trajectory Management. In 2023 24th IEEE International Conference on Mobile Data Management (MDM) . IEEE, 168–171

work page 2023
[26]

Tun-Hao You, Wen-Chih Peng, and Wang-Chien Lee. 2007. Protecting Moving Trajectories with Dummies. In 2007 International Conference on Mobile Data Management. 278–282. https://doi.org/10.1109/MDM.2007.58

work page doi:10.1109/mdm.2007.58 2007
[27]

Ming Yu, Jiecong Shi, Cuihong Xue, Xiaoke Hao, and Gang Yan. 2024. A review of single image super-resolution reconstruction based on deep learning.Multimedia Tools and Applications 83, 18 (2024), 55921–55962

work page 2024
[28]

Jing Yuan, Yu Zheng, Chengyang Zhang, Xing Xie, and Guang-Zhong Sun. 2010. An interactive-voting based map matching algorithm. In 2010 Eleventh interna- tional conference on mobile data management . IEEE, 43–52

work page 2010
[29]

Yu Zheng, Hao Fu, Xing Xie, Wei-Ying Ma, and Quannan Li. 2011. Ge- olife GPS trajectory dataset - User Guide (geolife gps trajectories 1.1 ed.). https://www.microsoft.com/en-us/research/publication/geolife-gps- trajectory-dataset-user-guide/

work page 2011
[30]

Yuanshao Zhu, Yongchao Ye, Xiangyu Zhao, and James JQ Yu. 2023. Diffusion Model for GPS Trajectory Generation. arXiv preprint arXiv:2304.11582 (2023)

work page arXiv 2023

[1] [1]

Osman Abul, Francesco Bonchi, and Mirco Nanni. 2008. Never Walk Alone: Uncertainty for Anonymity in Moving Objects Databases. In 2008 IEEE 24th International Conference on Data Engineering . 376–385

work page 2008

[2] [2]

Chen Chen, Cewu Lu, Qixing Huang, Qiang Yang, Dimitrios Gunopulos, and Leonidas Guibas. 2016. City-Scale Map Creation and Updating using GPS Collec- tions. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowl- edge Discovery and Data Mining (KDD ’16). Association for Computing Machinery, New York, NY, USA, 1465–1474. https://doi.org/10....

work page doi:10.1145/2939672.2939833 2016

[3] [3]

Yumeki Goto, Tomoya Matsumoto, Hamada Rizk, Naoto Yanai, and Hirozumi Yamaguchi. 2023. Privacy-Preserving Taxi-Demand Prediction Using Federated Learning. In International Conference on Smart Computing . IEEE

work page 2023

[4] [4]

Hemkumar, S

D. Hemkumar, S. Ravichandra, and D.V.L.N. Somayajulu. 2020. Impact of prior knowledge on privacy leakage in trajectory data publishing. Engineering Science and Technology, an International Journal 23, 6 (2020), 1291–1300

work page 2020

[5] [5]

Hongbo Jiang, Jie Li, Ping Zhao, Fanzi Zeng, Zhu Xiao, and Arun Iyengar. 2021. Location Privacy-Preserving Mechanisms in Location-Based Services: A Com- prehensive Survey. ACM Comput. Surv. 54, 1, Article 4 (jan 2021), 36 pages

work page 2021

[6] [6]

Kaifeng Jiang, Dongxu Shao, Stéphane Bressan, Thomas Kister, and Kian-Lee Tan. 2013. Publishing Trajectories with Differential Privacy Guarantees. In Pro- ceedings of the 25th International Conference on Scientific and Statistical Database Management (SSDBM). Association for Computing Machinery, New York, NY, USA, Article 12, 12 pages. https://doi.org/10....

work page doi:10.1145/2484838.2484846 2013

[7] [7]

Yang Li, Yangyan Li, Dimitrios Gunopulos, and Leonidas Guibas. 2016. Knowledge-based trajectory completion from sparse GPS samples. In Proceedings of the 24th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems. 1–10

work page 2016

[8] [8]

Yin Lou, Chengyang Zhang, Yu Zheng, Xing Xie, Wei Wang, and Yan Huang

work page

[9] [9]

In Proceedings of the 17th ACM SIGSPATIAL international conference on advances in geographic information systems

Map-matching for low-sampling-rate GPS trajectories. In Proceedings of the 17th ACM SIGSPATIAL international conference on advances in geographic information systems. 352–361

work page

[10] [10]

Antonios Makris, Camila Leite da Silva, Vania Bogorny, Luis Otavio Alvares, Jose Antonio Macedo, and Konstantinos Tserpes. 2021. Evaluating the effect of compressing algorithms for trajectory similarity and classification problems. Geoinformatica 25, 4 (oct 2021), 679–711

work page 2021

[11] [11]

Olsen, Jeong-Hyon Hwang, Catherine T

Jonathan Muckell, Paul W. Olsen, Jeong-Hyon Hwang, Catherine T. Lawson, and S. S. Ravi. 2014. Compression of trajectory data: a comprehensive evaluation and new approach. Geoinformatica 18, 3 (jul 2014), 435–460

work page 2014

[12] [12]

Mashaal Musleh and Mohamed F Mokbel. 2023. Kamel: A Scalable BERT-based System for Trajectory Imputation. Proceedings of the VLDB Endowment 17, 3 (2023), 525–538

work page 2023

[13] [13]

Paul Newson and John Krumm. 2009. Hidden Markov map matching through noise and sparseness. In Proceedings of the 17th ACM SIGSPATIAL international conference on advances in geographic information systems . 336–343

work page 2009

[14] [14]

Ren Ozeki, Haruki Yonekura, Aidana Baimbetova, Hamada Rizk, and Hirozumi Yamaguchi. 2023. One Model Fits All: Cross-Region Taxi-Demand Forecasting. In Proceedings of the 31st ACM International Conference on Advances in Geographic Information Systems (SIGSPATIAL ’23). Association for Computing Machinery, New York, NY, USA, Article 101, 4 pages. https://doi...

work page doi:10.1145/3589132.3625651 2023

[15] [15]

Ren Ozeki, Haruki Yonekura, Hamada Rizk, and Hirozumi Yamaguchi. 2022. Sharing without caring: privacy protection of users’ spatio-temporal data without compromise on utility. In Proceedings of the 30th International Conference on Advances in Geographic Information Systems . 1–2

work page 2022

[16] [16]

Ren Ozeki, Haruki Yonekura, Hamada Rizk, and Hirozumi Yamaguchi. 2023. Bal- ancing Privacy and Utility of Spatio-Temporal Data for Taxi-Demand Prediction. In The 24th IEEE International Conference on Mobile Data Management . IEEE

work page 2023

[17] [17]

Ren Ozeki, Haruki Yonekura, Hamada Rizk, and Hirozumi Yamaguchi. 2023. Balancing privacy and utility of spatio-temporal data for taxi-demand prediction. In 2023 24th IEEE International Conference on Mobile Data Management (MDM) . IEEE, 215–220

work page 2023

[18] [18]

Ren Ozeki, Haruki Yonekura, Hamada Rizk, and Hirozumi Yamaguchi. 2024. Privacy-Preserved Taxi Demand Prediction System Utilizing Distributed Data. In Proceedings of the 32nd ACM International Conference on Advances in Geographic Information Systems (SIGSPATIAL ’24). Association for Computing Machinery, to appear

work page 2024

[19] [19]

Qin, Yongli Ren, Wei Shao, Brennan Lake, Filippo Privitera, and Flora D

Kyle K. Qin, Yongli Ren, Wei Shao, Brennan Lake, Filippo Privitera, and Flora D. Salim. 2023. Multiple-level Point Embedding for Solving Human Trajectory Imputation with Prediction. ACM Trans. Spatial Algorithms Syst. 9, 2, Article 15 (may 2023), 22 pages. https://doi.org/10.1145/3582427

work page doi:10.1145/3582427 2023

[20] [20]

Jinmeng Rao, Song Gao, Yuhao Kang, and Qunying Huang. 2020. LSTM-TrajGAN: A Deep Learning Approach to Trajectory Privacy Protection. https://doi.org/10. 48550/ARXIV.2006.10521

work page arXiv 2020

[21] [21]

Sijie Ruan, Cheng Long, Jie Bao, Chunyang Li, Zisheng Yu, Ruiyuan Li, Yuxuan Liang, Tianfu He, and Yu Zheng. 2020. Learning to Generate Maps from Trajec- tories. Proceedings of the AAAI Conference on Artificial Intelligence 34, 01 (Apr. 2020), 890–897. https://ojs.aaai.org/index.php/AAAI/article/view/5435

work page 2020

[22] [22]

Akiyoshi Suzuki, Mayu Iwata, Yuki Arase, Takahiro Hara, Xing Xie, and Shojiro Nishio. 2010. A User Location Anonymization Method for Location Based Ser- vices in a Real Environment. In Proceedings of the 18th SIGSPATIAL International Conference on Advances in Geographic Information Systems (GIS ’10) . Association for Computing Machinery, New York, NY, USA...

work page 2010

[23] [23]

Jingyuan Wang, Ning Wu, Xinxi Lu, Wayne Xin Zhao, and Kai Feng. 2021. Deep Trajectory Recovery with Fine-Grained Calibration using Kalman Filter. IEEE Transactions on Knowledge and Data Engineering 33, 3 (2021), 921–934

work page 2021

[24] [24]

Szymon Woźniak and Piotr Szymański. 2021. Hex2vec: Context-aware embedding h3 hexagons with openstreetmap tags. In Proceedings of the 4th ACM SIGSPATIAL International Workshop on AI for Geographic Knowledge Discovery . 61–71

work page 2021

[25] [25]

Haruki Yonekura, Ren Ozeki, Hamada Rizk, and Hirozumi Yamaguchi. 2023. STM-A Privacy-Enhanced Solution for Spatio-Temporal Trajectory Management. In 2023 24th IEEE International Conference on Mobile Data Management (MDM) . IEEE, 168–171

work page 2023

[26] [26]

Tun-Hao You, Wen-Chih Peng, and Wang-Chien Lee. 2007. Protecting Moving Trajectories with Dummies. In 2007 International Conference on Mobile Data Management. 278–282. https://doi.org/10.1109/MDM.2007.58

work page doi:10.1109/mdm.2007.58 2007

[27] [27]

Ming Yu, Jiecong Shi, Cuihong Xue, Xiaoke Hao, and Gang Yan. 2024. A review of single image super-resolution reconstruction based on deep learning.Multimedia Tools and Applications 83, 18 (2024), 55921–55962

work page 2024

[28] [28]

Jing Yuan, Yu Zheng, Chengyang Zhang, Xing Xie, and Guang-Zhong Sun. 2010. An interactive-voting based map matching algorithm. In 2010 Eleventh interna- tional conference on mobile data management . IEEE, 43–52

work page 2010

[29] [29]

Yu Zheng, Hao Fu, Xing Xie, Wei-Ying Ma, and Quannan Li. 2011. Ge- olife GPS trajectory dataset - User Guide (geolife gps trajectories 1.1 ed.). https://www.microsoft.com/en-us/research/publication/geolife-gps- trajectory-dataset-user-guide/

work page 2011

[30] [30]

Yuanshao Zhu, Yongchao Ye, Xiangyu Zhao, and James JQ Yu. 2023. Diffusion Model for GPS Trajectory Generation. arXiv preprint arXiv:2304.11582 (2023)

work page arXiv 2023