Hui He

Dr. Hui He


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
7 April 2026 in Neural Processing Letters
Verified
Journal Article
Open access
PDF file
KWKefeng Wang
Kefeng Wang
JJJi Lin Jiang
Ji Lin Jiang
HHHui He
Hui He
JCJiajia Chen
Jiajia Chen
Accurately predicting the spatiotemporal distribution of traffic accident risk is crucial for proactive safety management and intelligent transportation systems. However, the inherent complex spatial dependencies and dynamic temporal patterns in traffic data, coupled with severe class imbalance in accident severity categories, pose significant challenges to conventional prediction models. This study proposes a novel deep learning framework that integrates Graph Neural Networks (GNN) and Long Short-Term Memory (LSTM) networks to address these challenges. The GNN module effectively captures complex spatial dependencies between geographical regions by modeling the traffic network as a graph structure, while the LSTM module learns the temporal dynamics of accident occurrences. To tackle the critical issue of data imbalance caused by scarce severe accident samples, this study employs the Adaptive Synthetic Sampling (ADASYN) technique at the data level to balance sample distributions, and at the model level, utilizes the Focal Loss function to focus the training process on hard-to-classify minority class instances. Furthermore, the model architecture incorporates TransformerConv layers, bidirectional LSTM, and multi-head attention mechanisms to capture more nuanced spatiotemporal features. Experimental evaluation on the large-scale US-Accidents dataset (2016–2021) demonstrates that the proposed GNN-LSTM fusion model significantly outperforms existing methods across multiple key metrics. On the test set, the model achieved a weighted F1-score of 0.7923 ± 0.0123 and a macro-average F1-score of 0.6587 ± 0.0156, representing improvements of 1.12 and 3.53 percentage points, respectively, over the strongest baseline model Graph WaveNet. In addressing class imbalance, the model attained an AUC-PR of 0.7124 ± 0.0098 and a Cohen’s Kappa of 0.6012 ± 0.0134, both significantly superior to conventional approaches. Ablation studies further validated the effectiveness of key components, with the GNN module contributing 13.8% to the performance gain, while ADASYN and Focal Loss contributed 8.7 and 6.5%, respectively. This study provides a robust and efficient solution for traffic risk prediction, offering a scientific foundation for traffic management authorities to develop precise early-warning systems and targeted intervention strategies.

7 August 2026 in Scientific Reports
Verified
Journal Article
Open access
PDF file
KWKefeng Wang
Kefeng Wang
HHHui He
Hui He
JJJilin Jiang
Jilin Jiang
XHXinyu He
Xinyu He
High-resolution vehicle trajectory data collected by drone-based traffic sensing systems support detailed microscopic traffic analysis, but they are often affected by missing values, measurement noise, and tracking errors. This paper proposes TPC-STGAN for urban vehicle trajectory imputation and denoising. The framework begins with trajectory prior calibration (TPC), which constructs a feature-wise Calibrated Prior from the corrupted trajectory and the observed-value mask. This prior serves as a structured residual anchor for a lightweight spatiotemporal generator based on dense projection, multi-head self-attention, bidirectional GRU sequence modeling, and a time-step sigmoid correction gate. The generator learns gated residual corrections around the prior, while a CNN-based discriminator encourages reconstructed trajectories to remain close to the distribution of clean trajectories. A feature-weighted mean squared error (FWMSE) reconstruction loss is used to reflect the different roles of position, speed, acceleration, and time variables. Experiments on the pNEUMA drone trajectory dataset are conducted under Gaussian+Random, Laplace+Random, and Drift+Block corruption protocols. In the main Gaussian+Random setting, TPC-STGAN achieves an RMSE of 0.0696 and an MAE of 0.0512, reducing corrupted-input RMSE by 86.47%. Under Drift+Block corruption, MovingAvg+Interp obtains the lowest overall RMSE, while TPC-STGAN ranks second overall and achieves the lowest RMSE among neural models. Ablation, sensitivity, robustness, downstream prediction, and deployment analyses further characterize component contributions, behavior across corruption types, and deployment cost.