Xuanshu Luo
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Xuanshu Luo has a total of 27 co-authors. Here are the top co-authors by number of shared publications:
- Martin Werner
Technical University of Munich
5 shared publications - Paul Walther
School of Engineering and Design, Technical University of Munich, Munich, Germany
4 shared publications - Hao Li
Department of Geography National University of Singapore Singapore Singapore
3 shared publications - Balthasar Teuscher
Technical University of Munich, Germany; TUM School of Engineering and Design, Department of Aerospace and Geodesy, Professorship of Big Geospatial Data Management
2 shared publications - Fabian Deuser
University of the Bundeswehr Munich, Munich, Germany
2 shared publications Martin Werner
Technical University of Munich, Germany; TUM School of Engineering and Design, Department of Aerospace and Geodesy, Professorship of Big Geospatial Data Management
2 shared publications
Latest Publications
- Open Access
- 4 August 2026
- Published by Copernicus GmbH in The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
- Vol. XLIX-B4-20, 217-222
Investigating Array Programming for Spatial Operations with Vector Geometries
Vector geometries are traditionally represented as entities of sequences of coordinate structures. With advancements in hardware and software for data analytics, a preference for columnar data layouts arose. This paper examines array programming for spatial operations to evaluate the potential performance benefits of modern computing architectures and emerging spatial data encodings. Evaluating selected operations, such as distance calculation, extent extraction, and affine transformations, indicates similar or improved performance for geometries with columnar coordinate layouts. By leveraging modern compiler infrastructure, we further demonstrate that advanced hardware features in commodity hardware, such as vector instructions, are becoming accessible without specialized code. The performance comparison with established, widely used geospatial and computational libraries reveals significant untapped potential for increasing the efficiency of spatial computing.
- 4 August 2026
- Published by Copernicus GmbH in The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
- Vol. XLIX-B4-20, 217-222
Open AccessInvestigating Array Programming for Spatial Operations with Vector Geometries
Vector geometries are traditionally represented as entities of sequences of coordinate structures. With advancements in hardware and software for data analytics, a preference for columnar data layouts arose. This paper examines array programming for spatial operations to evaluate the potential performance benefits of modern computing architectures and emerging spatial data encodings. Evaluating selected operations, such as distance calculation, extent extraction, and affine transformations, indicates similar or improved performance for geometries with columnar coordinate layouts. By leveraging modern compiler infrastructure, we further demonstrate that advanced hardware features in commodity hardware, such as vector instructions, are becoming accessible without specialized code. The performance comparison with established, widely used geospatial and computational libraries reveals significant untapped potential for increasing the efficiency of spatial computing.
- Open Access
- 3 July 2026
- Published by Copernicus GmbH in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
- Vol. XI-2-2026, 697-704
Quantization-Aware Training for Efficient Object Detection on FPGAs: Case Studies
Deploying object detection models for resource-constrained remote sensing applications necessitates on-board model inference capabilities. While Field Programmable Gate Arrays (FPGAs) offer massive parallelism as energy-efficient hardware platforms, model quantization remains essential to further balance computational efficiency with detection accuracy. Compared to post-training quantization methods that involve multiple-stage development with consistent dependency on domain datasets, quantization-aware training (QAT) integrates quantization constraints into training, providing a simpler pipeline for model compression. However, QAT introduces quantization errors to which smaller objects are more vulnerable. To address this issue, we propose object-scale-aware (OSA) regularization that amplifies quantization error penalties for smaller targets. Our approach is validated through two case studies: bird detection at airports and aerial-view building detection. We perform 8-bit QAT on YOLOX series models using the MVA2023 dataset and the Bavarian Building Dataset for the respective studies. Our method achieves up to 50.2 times inference acceleration with minimal accuracy loss on Xilinx Kria KV260 FPGAs compared to full-precision models. The ablation study and detection examples further demonstrate the effectiveness of OSA regularization in small object detection.
- 3 July 2026
- Published by Copernicus GmbH in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
- Vol. XI-2-2026, 697-704
Open AccessQuantization-Aware Training for Efficient Object Detection on FPGAs: Case Studies
Deploying object detection models for resource-constrained remote sensing applications necessitates on-board model inference capabilities. While Field Programmable Gate Arrays (FPGAs) offer massive parallelism as energy-efficient hardware platforms, model quantization remains essential to further balance computational efficiency with detection accuracy. Compared to post-training quantization methods that involve multiple-stage development with consistent dependency on domain datasets, quantization-aware training (QAT) integrates quantization constraints into training, providing a simpler pipeline for model compression. However, QAT introduces quantization errors to which smaller objects are more vulnerable. To address this issue, we propose object-scale-aware (OSA) regularization that amplifies quantization error penalties for smaller targets. Our approach is validated through two case studies: bird detection at airports and aerial-view building detection. We perform 8-bit QAT on YOLOX series models using the MVA2023 dataset and the Bavarian Building Dataset for the respective studies. Our method achieves up to 50.2 times inference acceleration with minimal accuracy loss on Xilinx Kria KV260 FPGAs compared to full-precision models. The ablation study and detection examples further demonstrate the effectiveness of OSA regularization in small object detection.
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 19, 24039-24055
SAGINGeo: A Space–Aerial–Ground Integrated Framework for VGI Geolocalization in Multidisaster Scenarios
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 19, 24039-24055
Open AccessSAGINGeo: A Space–Aerial–Ground Integrated Framework for VGI Geolocalization in Multidisaster Scenarios
- Open Access
- 3 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 1238-1241
Human Mobility Prediction with Multi-Task Curriculum Training
- 3 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 1238-1241
Open AccessHuman Mobility Prediction with Multi-Task Curriculum Training
- Top Cited PapersOpen Access
- 31 January 2025
- Published by Elsevier in ISPRS Journal of Photogrammetry and Remote Sensing
- Vol. 220, 841-854
Cross-view geolocalization and disaster mapping with street-view and VHR satellite imagery: A case study of Hurricane IAN
- 31 January 2025
- Published by Elsevier in ISPRS Journal of Photogrammetry and Remote Sensing
- Vol. 220, 841-854
Top Cited PapersOpen AccessCross-view geolocalization and disaster mapping with street-view and VHR satellite imagery: A case study of Hurricane IAN
- Open Access
- 29 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 31-34
TraBiMap: Reducing Privacy Concerns in Trajectory Analysis with Randomized Data Representations
Personal trajectory data, increasingly acquired by more and more GNSS-enabled devices, is currently underutilized as privacy concerns prohibit its comprehensive excavation to improve reaction and planning opportunities for communities and individuals. However, due to their distinctive data structure, existing data anonymization methods are difficult to apply directly to trajectories. Therefore, inspired by recent probabilistic representations of geographic information, we present TraBiMaps, a randomized data structure for trajectories based on Bloom Filters using cryptographic hash functions that can efficiently store and evaluate rasterized trajectories with a high level of individual privacy. We further provide a preliminary privacy analysis of TraBiMap and pose additional research questions in this field.
- 29 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 31-34
Open AccessTraBiMap: Reducing Privacy Concerns in Trajectory Analysis with Randomized Data Representations
Personal trajectory data, increasingly acquired by more and more GNSS-enabled devices, is currently underutilized as privacy concerns prohibit its comprehensive excavation to improve reaction and planning opportunities for communities and individuals. However, due to their distinctive data structure, existing data anonymization methods are difficult to apply directly to trajectories. Therefore, inspired by recent probabilistic representations of geographic information, we present TraBiMaps, a randomized data structure for trajectories based on Bloom Filters using cryptographic hash functions that can efficiently store and evaluate rasterized trajectories with a high level of individual privacy. We further provide a preliminary privacy analysis of TraBiMap and pose additional research questions in this field.
- Open Access
- 13 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-4
Exploring GeoAI Methods for Supraglacial Lake Mapping on Greenland Ice Sheet
The ACM SIGSPATIAL Cup 2023 proposed the challenge to identify and map supraglacial lakes in Greenland in satellite imagery. The peculiarities of supraglacial lakes pose a hard problem for semantic segmentation and object detection tasks because the definition of a lake is ill-fitted to the inner workings of such approaches. For example, lakes are often covered by ice and snow and narrow streams can connect distinct lakes, which is not directly translatable to the semantic segmentation of water. It is also not well-posed for object detection, especially the identity relation - what is a lake, what is not (yet) a lake, and what are two lakes is challenging. In this context, we worked on adapting semantic segmentation using the Segment Anything Model and instance segmentation using Mask R-CNN to the setting. The latter ended up superior in our own evaluation and even got ranked second among all participants. We are proud that our approach has led to competitive performance. The source code is available from https://github.com/tum-bgd/GISCup23.
- 13 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-4
Open AccessExploring GeoAI Methods for Supraglacial Lake Mapping on Greenland Ice Sheet
The ACM SIGSPATIAL Cup 2023 proposed the challenge to identify and map supraglacial lakes in Greenland in satellite imagery. The peculiarities of supraglacial lakes pose a hard problem for semantic segmentation and object detection tasks because the definition of a lake is ill-fitted to the inner workings of such approaches. For example, lakes are often covered by ice and snow and narrow streams can connect distinct lakes, which is not directly translatable to the semantic segmentation of water. It is also not well-posed for object detection, especially the identity relation - what is a lake, what is not (yet) a lake, and what are two lakes is challenging. In this context, we worked on adapting semantic segmentation using the Segment Anything Model and instance segmentation using Mask R-CNN to the setting. The latter ended up superior in our own evaluation and even got ranked second among all participants. We are proud that our approach has led to competitive performance. The source code is available from https://github.com/tum-bgd/GISCup23.
- Open Access
A Codeword-Independent Localization Technique for Reconfigurable Intelligent Surface Enhanced Environments Using Adversarial Learning
Reconfigurable Intelligent Surfaces (RISs) not only enable software-defined radio in modern wireless communication networks but also have the potential to be utilized for localization. Most previous works used channel matrices to calculate locations, requiring extensive field measurements, which leads to rapidly growing complexity. Although a few studies have designed fingerprint-based systems, they are only feasible under an unrealistic assumption that the RIS will be deployed only for localization purposes. Additionally, all these methods utilize RIS codewords for location inference, inducing considerable communication burdens. In this paper, we propose a new localization technique for RIS-enhanced environments that does not require RIS codewords for online location inference. Our proposed approach extracts codeword-independent representations of fingerprints using a domain adversarial neural network. We evaluated our solution using the DeepMIMO dataset. Due to the lack of results from other studies, for fair comparisons, we define oracle and baseline cases, which are the theoretical upper and lower bounds of our system, respectively. In all experiments, our proposed solution performed much more similarly to the oracle cases than the baseline cases, demonstrating the effectiveness and robustness of our method.
Open AccessA Codeword-Independent Localization Technique for Reconfigurable Intelligent Surface Enhanced Environments Using Adversarial Learning
Reconfigurable Intelligent Surfaces (RISs) not only enable software-defined radio in modern wireless communication networks but also have the potential to be utilized for localization. Most previous works used channel matrices to calculate locations, requiring extensive field measurements, which leads to rapidly growing complexity. Although a few studies have designed fingerprint-based systems, they are only feasible under an unrealistic assumption that the RIS will be deployed only for localization purposes. Additionally, all these methods utilize RIS codewords for location inference, inducing considerable communication burdens. In this paper, we propose a new localization technique for RIS-enhanced environments that does not require RIS codewords for online location inference. Our proposed approach extracts codeword-independent representations of fingerprints using a domain adversarial neural network. We evaluated our solution using the DeepMIMO dataset. Due to the lack of results from other studies, for fair comparisons, we define oracle and baseline cases, which are the theoretical upper and lower bounds of our system, respectively. In all experiments, our proposed solution performed much more similarly to the oracle cases than the baseline cases, demonstrating the effectiveness and robustness of our method.
- 17 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 23 (1) , 895-908
Touchscreens Can Reveal User Identity: Capacitive Plethysmogram-Based Biometrics
Biometrics are widely used for user identification/authentication, but the fact has rarely been noticed that general capacitive touchscreens can reveal user identities by touch signals. This paper proposes a new biometric method with inherent liveness detection for reliable user recognition based on the cardiac signal captured by the capacitive touchscreen, namely Capacitive Plethysmogram (CPG). And a systematic framework is designed for CPG collection, processing, and exploitation to identify users. Specifically, since the finger usually forms capacitors with multiple sensing electrodes during touching, we can extract several CPG signals simultaneously from the screen output. Then we propose a series of preprocessing algorithms to filter CPG for signal quality enhancement. Finally, to further leverage filtered CPG signals and extract efficient features for identifying users, we build an encoder based on 3D attention CNN and metric learning. Experimental results demonstrate that the proposed method can achieve an average accuracy of 96.73%, FAR of 3.03%, and FRR of 7.35% in the laboratory environment, which reveals the potential of CPG for user privacy protection and data security on various devices laced with capacitive touchscreens.
- 17 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 23 (1) , 895-908
Touchscreens Can Reveal User Identity: Capacitive Plethysmogram-Based Biometrics
Biometrics are widely used for user identification/authentication, but the fact has rarely been noticed that general capacitive touchscreens can reveal user identities by touch signals. This paper proposes a new biometric method with inherent liveness detection for reliable user recognition based on the cardiac signal captured by the capacitive touchscreen, namely Capacitive Plethysmogram (CPG). And a systematic framework is designed for CPG collection, processing, and exploitation to identify users. Specifically, since the finger usually forms capacitors with multiple sensing electrodes during touching, we can extract several CPG signals simultaneously from the screen output. Then we propose a series of preprocessing algorithms to filter CPG for signal quality enhancement. Finally, to further leverage filtered CPG signals and extract efficient features for identifying users, we build an encoder based on 3D attention CNN and metric learning. Experimental results demonstrate that the proposed method can achieve an average accuracy of 96.73%, FAR of 3.03%, and FRR of 7.35% in the laboratory environment, which reveals the potential of CPG for user privacy protection and data security on various devices laced with capacitive touchscreens.
- 5 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
A Geometric Deep Learning Framework for Accurate Indoor Localization
Recent advances in (deep) machine learning offer new opportunities to solve indoor fingerprint-based localization problems. However, the majority of localization solutions employing popular machine learning models, such as k-nearest neighbors ( $k$ -NN), support vector machine (SVM), multi-layer perceptron (MLP), and convolutional neural network (CNN), do not sufficiently realize inability of these models to fully represent the non-Euclidean nature of fingerprint data, which consequently degrades their performance. In this paper, we first explain how these commonly-used models fail to effectively encode the fingerprint data due to their assumption (or lack of it) regarding fingerprints and/or geometric and topology information hidden within the RSSI measurements. Based on this, we provide our motivation to use geometric deep learning for indoor fingerprint-based localization. We then present a systematic approach to transform fingerprints into graphs, accounting for the co-existence of multiple radio frequency signal technologies. Finally, we present our localization approach based on a GraphSAGE estimator. Through extensive performance evaluation, using two different case studies (datasets), we show to what extent our proposed localization approach improves upon the state-of-the-art localization solutions. We also conclude that the best configuration of our approach requires both the edge features in the graphs and the pooling aggregator in the GraphSAGE model.- 5 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
A Geometric Deep Learning Framework for Accurate Indoor Localization
Recent advances in (deep) machine learning offer new opportunities to solve indoor fingerprint-based localization problems. However, the majority of localization solutions employing popular machine learning models, such as k-nearest neighbors ( $k$ -NN), support vector machine (SVM), multi-layer perceptron (MLP), and convolutional neural network (CNN), do not sufficiently realize inability of these models to fully represent the non-Euclidean nature of fingerprint data, which consequently degrades their performance. In this paper, we first explain how these commonly-used models fail to effectively encode the fingerprint data due to their assumption (or lack of it) regarding fingerprints and/or geometric and topology information hidden within the RSSI measurements. Based on this, we provide our motivation to use geometric deep learning for indoor fingerprint-based localization. We then present a systematic approach to transform fingerprints into graphs, accounting for the co-existence of multiple radio frequency signal technologies. Finally, we present our localization approach based on a GraphSAGE estimator. Through extensive performance evaluation, using two different case studies (datasets), we show to what extent our proposed localization approach improves upon the state-of-the-art localization solutions. We also conclude that the best configuration of our approach requires both the edge features in the graphs and the pooling aggregator in the GraphSAGE model.