Zijun Yu

Zijun Yu

Auto-generated profile, based on publicly available information

Zijun Yu

Shanghai Maritime University

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

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

Representative Publications

View All Publications

25 August 2023
Verified
Conference Paper
Open access
SZShenjie Zou
Shenjie Zou
JLJin Liu
Jin Liu
XZXiliang Zhang
Xiliang Zhang
ZYZijun Yu
Zijun Yu
BHBing Han
Bing Han
JZJiamao Zhi
Jiamao Zhi
In maritime transportation, accurate prediction of AIS-based vessel trajectory has a significant impact on collision-free vessel trajectory prediction. Due to the dense and highly dynamic marine navigation traffic, it is a great challenge to predict the vessels' trajectories. Currently, the mainstream trajectory prediction methods are based on recurrent neural networks ignoring spatial-temporal interaction between vessels. In order to fully learn the structural features of complex trajectories and the interaction features between motion targets in complex scenarios, we propose a Dynamic Spatial-Temporal Refinement Network (DSTNet). Our model introduces the states refinement module and Temporal Transformer. Specifically, the vessel attention factor and motion gate are integrated to implement the interaction of vessel spatial trajectory information. Moreover, Temporal Transformer is used to obtain global traffic characteristics, extracting the impact of traffic-agents beyond the specified spatial limits and integrate spatio-temporal data. The results of our proposed method on a real-world Automatic Identification System (AIS) dataset show 18.8% and 25% improvement in ADE and FDE metrics compared to other methods.


8 July 2025 in IEEE Transactions on Computational Social Systems
Verified
Journal Article
Closed access
ZYZijun Yu
Zijun Yu
JLJin Liu
Jin Liu
PGPeizhu Gong
Peizhu Gong
XLXingye Li
Xingye Li
LWLai Wei
Lai Wei
WZWu Zhongdai
Wu Zhongdai
The key to address long-term person reidentification (Re-ID) in videos is to extract invariant spatio-temporal features (ISTF), which can be broadly categorized into two forms: 1) clothes-irrelevant appearance such as facial characteristics and body shape; 2) identity-distinctive motion such as posture and gait. However, existing studies mainly focus on mining either appearance- or motion-based features in the sequences without sufficient utilization of the ISTF. In this article, we propose an Appearance and gait-based tango network (AGT-Net) to comprehensively mine ISTF from appearance details and gait motions. Specifically, on the appearance detail branch, we introduce an invariance-aware video Swin transformer (IA-VST) to extract clothes-irrelevant appearance from the original RGB sequences. On the gait motion branch, we propose a motion-sensitive gait feature extractor (MS-GFE) to learn identity-distinctive motion from the pre-processed gait sequences. In addition, a score-level fusion strategy is introduced to integrate information from the two streams for prediction. Besides, since there is a lack of publicly available datasets, we propose a style-transferring synthetic long-term Video Re-ID (STYLE-VID) dataset, particularly for long-term Re-ID. Extensive experiments demonstrate that AGT-Net not only outperforms the state-of-the-art methods by up to 1.5% mAP on STYLE-VID, but also achieves comparable performance to other models on the traditional short-term Re-ID dataset MARS.