Xiliang Zhang

Xiliang Zhang

Auto-generated profile, based on publicly available information

Xiliang Zhang

College of Information Engineering, Shanghai Maritime University, Shanghai, China

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
26 September 2023 in Ocean Engineering
Verified
Journal Article
Closed access
XZXiliang Zhang
Xiliang Zhang
JLJin Liu
Jin Liu
PGPeizhu Gong
Peizhu Gong
CCChengcheng Chen
Chengcheng Chen
BHBing Han
Bing Han
WZWu Zhongdai
Wu Zhongdai

24 February 2023 in Complex & Intelligent Systems
Verified
Journal Article
Open access
PDF file
SHShixing Han
Shixing Han
JLJin Liu
Jin Liu
JZJinyingming Zhang
Jinyingming Zhang
PGPeizhu Gong
Peizhu Gong
XZXiliang Zhang
Xiliang Zhang
HHHuihua He
Huihua He
Dense video captioning (DVC) aims at generating description for each scene in a video. Despite attractive progress for this task, previous works usually only concentrate on exploiting visual features while neglecting audio information in the video, resulting in inaccurate scene event location. In this article, we propose a novel DVC model named CMCR, which is mainly composed of a cross-modal processing (CM) module and a commonsense reasoning (CR) module. CM utilizes a cross-modal attention mechanism to encode data in different modalities. An event refactoring algorithm is proposed to deal with inaccurate event localization caused by overlapping events. Besides, a shared encoder is utilized to reduce model redundancy. CR optimizes the logic of generated captions with both heterogeneous prior knowledge and entities’ association reasoning achieved by building a knowledge-enhanced unbiased scene graph. Extensive experiments are conducted on ActivityNet Captions dataset, the results demonstrate that our model achieves better performance than state-of-the-art methods. To better understand the performance achieved by CMCR, we also apply ablation experiments to analyze the contributions of different modules.

6 September 2023 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
CCChengcheng Chen
Chengcheng Chen
WZWeiming Zeng
Weiming Zeng
XZXiliang Zhang
Xiliang Zhang
YZYuhao Zhou
Yuhao Zhou
In recent years, transformer-based networks, known for their ability to model long-range dependencies, have been widely used in downstream computer vision tasks, surpassing certain neural network architectures. However, transformer-based networks suffer from issues such as large parameter size, high computational complexity, and difficulties in extending spatial and channel features to the third or even higher orders, resulting in convergence challenges for small to medium-sized datasets and limited effectiveness in extracting high-order detailed features. In this paper, we propose a high-order spatial and channel controllable convolution module, named CS n , which can replace standard convolutions in any convolutional network. In the context of remote sensing small object detection, CS n demonstrates superior performance compared to Self-Attention structures embedded in neural networks. Moreover, it introduces long-range dependency relationships among pixels, similar to Self-Attention, and adopts a cascaded recursive approach to extend spatial and channel features to arbitrary higher orders without introducing significant additional computation. This extension captures crucial information from high-order spatial and channel dimensions, resulting in improved accuracy for small object detection. Additionally, we construct a novel, versatile CS n -(FPN+PAN) structure for object detection networks, referred to as CS n Net. Finally, our proposed model exhibits significant advantages in remote sensing detection when compared to state-of-the-art methods on publicly available SAR datasets (SSDD, HRSID) and optical remote sensing dataset (NWPU-10), achieving respective improvements of 3.4%, 4.5%, and 0.4% in mAP 50 compared to baseline models.

31 December 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
YLYuming Li
Yuming Li
JLJin Liu
Jin Liu
XLXingye Li
Xingye Li
XZXiliang Zhang
Xiliang Zhang
WZWu Zhongdai
Wu Zhongdai
BHBing Han
Bing Han
Recently, with the increasing adoption of synthetic aperture radar (SAR) ship detection methods on mobile platforms, the lightweighting of detection methods has become a research focus. Despite certain achievements, there are still several limitations: (1) Existing studies have mainly focused on reducing model complexity through shallow network structures. However, this approach frequently results in performance degradation, as they neglected a thorough investigation into achieving a better balance point between inference speed and detection accuracy. (2) Under the lightweight network structure, the rich edge features contained in SAR images, which are crucial for distinguishing ship targets from complex backgrounds, are often underutilized. To address these issues, we propose a novel lightweight detection method based on edge feature aware and fusion. Specifically, to effectively extract edge feature, we introduce an Edge FeatureAware (EFA) network that incorporates a multi-scale channel attention module. Furthermore, a lightweight feature fusion network, Filter-Pruned Bi-directional Feature Pyramid Network (FP-BiFPN), is carefully designed, which can not only suppresses background information but also accentuates ship targets. Finally, we propose a selective quantization algorithm based on a bit-width selection mechanism to reduce model memory usage without compromising performance. To validate the superiority of our proposed method, we conduct extensive experiments on multiple public datasets, achieving average accuracy scores of 94.2%, 97.6%, and 97.7% on the HRSID, SAR-Ship-Dataset, and SSDD, respectively, with a model parameter size of only 3.36M, and the fastest processing time for a single frame is 7.2ms.

12 November 2024 in Expert Systems with Applications
Verified
Journal Article
Closed access
XZXiliang Zhang
Xiliang Zhang
JLJin Liu
Jin Liu
CCChengcheng Chen
Chengcheng Chen
LWLai Wei
Lai Wei
WZWu Zhongdai
Wu Zhongdai
WDWenjuan Dai
Wenjuan Dai

16 May 2024 in IEEE Transactions on Intelligent Vehicles
Verified
Journal Article
Closed access
XZXiliang Zhang
Xiliang Zhang
JLJin Liu
Jin Liu
KCKejie Chen
Kejie Chen
PGPeizhu Gong
Peizhu Gong
YLYuxin Liu
Yuxin Liu
WZWu Zhongdai
Wu Zhongdai
Precise multi-vessel trajectory prediction (MVTP) in intelligent collision avoidance systems is a crucial yet challenging task due to the dynamic and continuous spatio-temporal dependencies between different vessels. However, most existing frameworks utilize graph structures with shared nodes to capture static correlations and employ sophisticated mechanisms to learn short-term trends hidden in trajectories, which fail to provide sufficient insight into the dynamic interactions and long-term patterns among vessels, limiting prediction accuracy and effectiveness. To address these pivotal issues, we propose a Dynamic Spatio-Temporal Graph Attention Network (DSTGAN) for MVTP, catering to intricate interactions modeling with long- and short-term motion pattern perception. In the spatial dimension, we present a vessel spatial graph encoder (VSGE) to learn dynamic interactions, which adaptively reconstructs the adjacency matrix from historical vessel data instead of constructing a fixed graph according to vessel proximity. Also, we introduce attribute-aware graph attention in VSGE to guide customized information propagation, and an adaptive neighbor sampling strategy to control the quality of vessel feature aggregation. For the temporal dimension, a vessel temporal graph fusion module (VTGF) is designed to enhance the perception of temporal dependencies at both local and global scales. To prevent information leakage, we propose masked time-aware attention, which automatically masks the features of future vessel nodes during attention computation. Furthermore, we incorporate a residual control block (RCB) to improve noise resilience and model convergence. Finally, the multi-vessel trajectory prediction module combines spatio-temporal features from VSGE and VTGF for subsequent prediction. Extensive experiments on three real-world Automatic Identification System (AIS) datasets demonstrate the superior performance of our method in comparison to state-of-the-art approaches.