Heshan Du
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
Heshan Du has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
Glen Hart
University of Nottingham
4 shared publications- Natasha Alechina
Open Universiteit, The Netherlands Utrecht University, The Netherlands
4 shared publications - Jianfeng Ren
School of Computer Science, University of Nottingham Ningbo China, Ningbo, China
3 shared publications Michael Jackson
University of Nottingham
3 shared publicationsNatasha Alechina
3 shared publications- Chenglin Yao
Changchun University, Changchun, Jilin, CHINA
2 shared publications
Latest Publications
- 26 October 2026
- Published by Wiley in Transactions in GIS
- Vol. 30 (7)
Assessing the Quality of OSM 3D Building Data Using 3D Geometric Comparison and Machine Learning
OpenStreetMap (OSM) is a well‐known platform for Volunteered Geographic Information. Most studies on OSM 3D building quality focus on a single aspect, such as footprint shape errors or attribute completeness. Current machine learning approaches typically analyze 2D footprints, with little attention paid to 3D characteristics like roof geometry or building height. To fill this gap, we propose an integrated framework that combines geometric comparative evaluation with machine‐learning prediction. In this framework, based on footprint Intersection over Union (IoU) and centroid distance, OSM buildings are matched with authoritative or externally validated 3D reference datasets selected for each study area, for example, from national mapping agencies. For the matched OSM buildings, the spatial accuracy is evaluated using location error, height error, building area error, and 3D Hausdorff distance error. For the unmatched buildings, machine learning models are used to predict their quality. We validated the proposed framework in six study areas across Asia and Europe: Beijing, Ningbo, Osaka, Yokohama, Amsterdam, and Almere. In total, the evaluation covers more than 115,000 OSM buildings. Compared with traditional methods, the proposed framework significantly improves facade error detection, evaluation completeness, and the granularity of quality classification, offering new insights for assessing and improving OSM data quality. The results produced in this study are available at: https://github.com/Longtao2002/OSM‐3D‐Building‐Quality‐Assessment .
- 26 October 2026
- Published by Wiley in Transactions in GIS
- Vol. 30 (7)
Assessing the Quality of OSM 3D Building Data Using 3D Geometric Comparison and Machine Learning
OpenStreetMap (OSM) is a well‐known platform for Volunteered Geographic Information. Most studies on OSM 3D building quality focus on a single aspect, such as footprint shape errors or attribute completeness. Current machine learning approaches typically analyze 2D footprints, with little attention paid to 3D characteristics like roof geometry or building height. To fill this gap, we propose an integrated framework that combines geometric comparative evaluation with machine‐learning prediction. In this framework, based on footprint Intersection over Union (IoU) and centroid distance, OSM buildings are matched with authoritative or externally validated 3D reference datasets selected for each study area, for example, from national mapping agencies. For the matched OSM buildings, the spatial accuracy is evaluated using location error, height error, building area error, and 3D Hausdorff distance error. For the unmatched buildings, machine learning models are used to predict their quality. We validated the proposed framework in six study areas across Asia and Europe: Beijing, Ningbo, Osaka, Yokohama, Amsterdam, and Almere. In total, the evaluation covers more than 115,000 OSM buildings. Compared with traditional methods, the proposed framework significantly improves facade error detection, evaluation completeness, and the granularity of quality classification, offering new insights for assessing and improving OSM data quality. The results produced in this study are available at: https://github.com/Longtao2002/OSM‐3D‐Building‐Quality‐Assessment .
- 24 June 2025
- Published by Springer Nature
- p. 286-300
Dual-Branch StarNet with Mutual Attention and U-Net Denoising for Simultaneously Recognizing Keywords and Speakers
- 24 June 2025
- Published by Springer Nature
- p. 286-300
Dual-Branch StarNet with Mutual Attention and U-Net Denoising for Simultaneously Recognizing Keywords and Speakers
- Open Access
- 17 July 2024
- Published by Elsevier in Expert Systems with Applications
- Vol. 255
Progressively-orthogonally-mapped EfficientNet for action recognition on time-range-Doppler signature
- 17 July 2024
- Published by Elsevier in Expert Systems with Applications
- Vol. 255
Open AccessProgressively-orthogonally-mapped EfficientNet for action recognition on time-range-Doppler signature
- 10 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 18 (15566013) , 4313-4328
Mask Attack Detection Using Vascular-Weighted Motion-Robust rPPG Signals
Detecting 3D mask attacks to a face recognition system is challenging. Although genuine faces and 3D face masks show significantly different remote photoplethysmography (rPPG) signals, rPPG-based face anti-spoofing methods often suffer from performance degradation due to unstable face alignment in the video sequence and weak rPPG signals. To enhance the rPPG signal in a motion-robust way, a landmark-anchored face stitching method is proposed to align the faces robustly and precisely at the pixel-wise level by using both SIFT keypoints and facial landmarks. To better encode the rPPG signal, a weighted spatial-temporal representation is proposed, which emphasizes the face regions with rich blood vessels. In addition, characteristics of rPPG signals in different color spaces are jointly utilized. To improve the generalization capability, a lightweight EfficientNet with a Gated Recurrent Unit (GRU) is designed to extract both spatial and temporal features from the rPPG spatial-temporal representation for classification. The proposed method is compared with the state-of-the-art methods on five benchmark datasets under both intra-dataset and cross-dataset evaluations. The proposed method shows a significant and consistent improvement in performance over other state-of-the-art rPPG-based methods for face spoofing detection.
- 10 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 18 (15566013) , 4313-4328
Mask Attack Detection Using Vascular-Weighted Motion-Robust rPPG Signals
Detecting 3D mask attacks to a face recognition system is challenging. Although genuine faces and 3D face masks show significantly different remote photoplethysmography (rPPG) signals, rPPG-based face anti-spoofing methods often suffer from performance degradation due to unstable face alignment in the video sequence and weak rPPG signals. To enhance the rPPG signal in a motion-robust way, a landmark-anchored face stitching method is proposed to align the faces robustly and precisely at the pixel-wise level by using both SIFT keypoints and facial landmarks. To better encode the rPPG signal, a weighted spatial-temporal representation is proposed, which emphasizes the face regions with rich blood vessels. In addition, characteristics of rPPG signals in different color spaces are jointly utilized. To improve the generalization capability, a lightweight EfficientNet with a Gated Recurrent Unit (GRU) is designed to extract both spatial and temporal features from the rPPG spatial-temporal representation for classification. The proposed method is compared with the state-of-the-art methods on five benchmark datasets under both intra-dataset and cross-dataset evaluations. The proposed method shows a significant and consistent improvement in performance over other state-of-the-art rPPG-based methods for face spoofing detection.
- Open Access
- 1 July 2020
- Published by International Joint Conferences on Artificial Intelligence
- p. 1695-1702
A Logic of Directions
We propose a logic of directions for points (LD) over 2D Euclidean space, which formalises primary direction relations east (E), west (W), and indeterminate east/west (Iew), north (N), south (S) and indeterminate north/south (Ins). We provide a sound and complete axiomatisation of it, and prove that its satisfiability problem is NP-complete.
- 1 July 2020
- Published by International Joint Conferences on Artificial Intelligence
- p. 1695-1702
Open AccessA Logic of Directions
We propose a logic of directions for points (LD) over 2D Euclidean space, which formalises primary direction relations east (E), west (W), and indeterminate east/west (Iew), north (N), south (S) and indeterminate north/south (Ins). We provide a sound and complete axiomatisation of it, and prove that its satisfiability problem is NP-complete.
- Open Access
- 23 September 2016
- Published by Springer Nature
- p. 30-37
An Ontology of Soil Properties and Processes
- 23 September 2016
- Published by Springer Nature
- p. 30-37
Open AccessAn Ontology of Soil Properties and Processes
- 31 August 2016
- Published by AI Access Foundation in Journal of Artificial Intelligence Research
- Vol. 56, 693-745
Qualitative Spatial Logics for Buffered Geometries
This paper describes a series of new qualitative spatial logics for checking consistency of sameAs and partOf matches between spatial objects from different geospatial datasets, especially from crowd-sourced datasets. Since geometries in crowd-sourced data are usually not very accurate or precise, we buffer geometries by a margin of error or a level of tolerance, and define spatial relations for buffered geometries. The spatial logics formalize the notions of `buffered equal' (intuitively corresponding to `possibly sameAs'), `buffered part of' (`possibly partOf'), `near' (`possibly connected') and `far' (`definitely disconnected'). A sound and complete axiomatisation of each logic is provided with respect to models based on metric spaces. For each of the logics, the satisfiability problem is shown to be NP-complete. Finally, we briefly describe how the logics are used in a system for generating and debugging matches between spatial objects, and report positive experimental evaluation results for the system.
- 31 August 2016
- Published by AI Access Foundation in Journal of Artificial Intelligence Research
- Vol. 56, 693-745
Qualitative Spatial Logics for Buffered Geometries
This paper describes a series of new qualitative spatial logics for checking consistency of sameAs and partOf matches between spatial objects from different geospatial datasets, especially from crowd-sourced datasets. Since geometries in crowd-sourced data are usually not very accurate or precise, we buffer geometries by a margin of error or a level of tolerance, and define spatial relations for buffered geometries. The spatial logics formalize the notions of `buffered equal' (intuitively corresponding to `possibly sameAs'), `buffered part of' (`possibly partOf'), `near' (`possibly connected') and `far' (`definitely disconnected'). A sound and complete axiomatisation of each logic is provided with respect to models based on metric spaces. For each of the logics, the satisfiability problem is shown to be NP-complete. Finally, we briefly describe how the logics are used in a system for generating and debugging matches between spatial objects, and report positive experimental evaluation results for the system.
- 9 May 2016
- Published by Wiley in Transactions in GIS
- Vol. 21 (2) , 406-427
A Method for Matching Crowd‐sourced and Authoritative Geospatial Data
- 9 May 2016
- Published by Wiley in Transactions in GIS
- Vol. 21 (2) , 406-427
A Method for Matching Crowd‐sourced and Authoritative Geospatial Data
- 1 May 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
A tool for matching crowd-sourced and authoritative geospatial data
A software tool MatchMaps was designed to facilitate matching between geospatial features held in different datasets, especially an authoritative dataset and a crowd-sourced dataset. MatchMaps is not fully automatic; it requires human interaction to resolve problematic matching cases. Previous studies have shown that this approach results in higher precision and recall than those achieved by fully automatic tools. In this study, we aim to evaluate MatchMaps with respect to the amount of human effort required for matching, and compare it with a fully manual matching process.
- 1 May 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
A tool for matching crowd-sourced and authoritative geospatial data
A software tool MatchMaps was designed to facilitate matching between geospatial features held in different datasets, especially an authoritative dataset and a crowd-sourced dataset. MatchMaps is not fully automatic; it requires human interaction to resolve problematic matching cases. Previous studies have shown that this approach results in higher precision and recall than those achieved by fully automatic tools. In this study, we aim to evaluate MatchMaps with respect to the amount of human effort required for matching, and compare it with a fully manual matching process.