Xia WU

Prof. Xia WU

orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

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

Representative Publications

View All Publications
5 December 2022 in iScience
Verified
Journal Article
Open access
CWChuntao Wu
Chuntao Wu
BHBeiyuan Hu
Beiyuan Hu
LWLei Wang
Lei Wang
XWXia WU
Xia WU
HGHaitao Gu
Haitao Gu
JLJiang Long
Jiang Long
Lipid metabolism is extensively reprogrammed in pancreatic ductal adenocarcinoma (PDAC). Stearoyl-coenzyme A desaturase (SCD) is a critical lipid regulator that was unexplored in PDAC. Here, we characterized the existence of cancer-associated fibroblasts (CAFs) with high SCD expression, and revealed them as an unfavorable prognostic factor. Therefore, primary CAFs and pancreatic cancer cells were harvested and genetically labeled. The mixture of CAFs and cancer cells were co-injected into scd-/-; prkdc-/-, or hIGF1/INS-expressing zebrafish to generate patient-derived xenograft models (zPDX). The models were aligned in 3D-printed chips for semi-automatic drug administration and high-throughput scanning. The results showed that chaperoning of the SCD-high CAFs significantly improved the drug resistance of pancreatic cancer cells against gemcitabine and cisplatin, while the administration of SCD inhibitors neutralized the protective effect. Our studies revealed the prognostic and therapeutic value of stromal SCD in PDAC, and proposed the application of zPDX model chips for drug testing.

20 March 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
JLJianhu Liu
Jianhu Liu
PWPing Wang
Ping Wang
XWXia WU
Xia WU
Vehicle-infrastructure cooperative perception enhances the perception capabilities of autonomous vehicles by facilitating the exchange of complementary information between vehicles and infrastructure. However, real-world environments often present challenges such as differences in sensor resolution and installation angles, which create a domain gap that complicates the integration of features from these two sources. This domain gap can hinder the overall performance of the perception system. To tackle this issue, we propose a novel vehicle–infrastructure cooperative perception network designed to effectively bridge the feature integration between vehicle and infrastructure sensors. Our approach includes a Multi-Scale Dynamic Feature Fusion Module designed to comprehensively integrate features from both vehicle and infrastructure across spatial and semantic dimensions. For feature fusion at each scale, we introduce the Multi-Source Dynamic Interaction Module (MSDI) and the Per-Point Self-Attention Module (PPSA). The MSDI dynamically adjusts the interaction between vehicle and infrastructure features based on environmental changes, generating enhanced interacting features. Subsequently, the PPSA aggregates these interacted features with the original vehicle–infrastructure features at the same spatial location. Additionally, we have constructed a real-world vehicle–infrastructure cooperative perception dataset, DZGSet, which includes multi-category annotations. Extensive experiments conducted on the DAIR-V2X and our self-collected DZGSet datasets demonstrate that our proposed method achieves Average Precision (AP) scores at IoU 0.5 of 0.780 and 0.652, and AP scores at IoU 0.7 of 0.632 and 0.493, respectively. These results indicate that our proposed method outperforms existing cooperative perception methods. Consequently, the proposed approach significantly improves the performance of cooperative perception, enabling more accurate and reliable autonomous vehicle operation.

2 January 2019 in Radio Science
Verified
Journal Article
Open access
PDF file
XWXia WU
Xia WU
In recent years, significant effort of both the scientific and industrial community is invested in research and development of multifrequency, multipolarization, and high‐resolution synthetic aperture radar (SAR) systems. One of the key issues is to realize efficient methods for imaging, reconstruction, and target detection of three‐dimensional (3‐D) electrically large‐scale objects. In this paper, an advanced multiple‐input multiple‐output (MIMO) SAR based on high‐resolution imaging algorithm was developed. Specifically, a reasonable sea surface model to quantify the influence of physical parameters in clutter was yielded as well. The SAR simulation system includes the following: (i) refined geometric‐electromagnetic modeling by Gmsh, (ii) a dynamic sea surface model, (iii) single‐input multiple‐output (SIMO) radar imaging formation by employing a fast backprojection algorithm, (iv) design of MIMO antenna arrays to achieve a good trade‐off between the main beam width and the side lobe level, (v) 3‐D SAR imaging algorithm on the basis of optimized physical optics for electromagnetic scattering computation, and (vi) hardware acceleration technique using graphics processing unit. Finally, numerical experiments indicate that 3‐D visualization of sphere, and examples of fighter aircraft and aircraft carrier are generated within 200 s, further proving correctness and effectiveness of the developed 3‐D MIMO SAR imaging algorithm.

6 April 2020
Verified
Conference Paper
Open access
PDF file
ZWZiao Wang
Ziao Wang
XZXiaofeng Zhang
Xiaofeng Zhang
XYXiaofei Yang
Xiaofei Yang
SWShaokai Wang
Shaokai Wang
XWXia WU
Xia WU
User portrait is an important technique which could be widely adopted in many real-world applications. It refers to the task of estimating user’s attributes containing their potential preferences and thus could be used to make better recommendations. Online dating Website is one of the most popular web applications which usually has more than 100 million users. Users of such Website are usually required to write down a brief self-introduction as well as fill in important attributes which are then used to make the friend recommendation. However, such data generally contains a good number of missing data or fake information. To cope with this challenge, this paper proposes a deep learning-based approach to generate user portrait merely from user’s self-introduction. Then, a multi-task deep neural network is adopted to simultaneously estimate five kinds of attributes, e.g., age, gender, education, salary and characteristics, from the input textual data. The experimental dataset is collected from one famous dating Website in China. Both the proposed approach as well as three benchmark approaches are implemented and the results are reported. The performance of the proposed approach is superior to the compared method with respect to accuracy.

24 June 2025 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
JLJianhu Liu
Jianhu Liu
XWXia WU
Xia WU
PWPing Wang
Ping Wang
Cooperative perception via Vehicle-to-Everything (V2X) wireless communication technology shares complementary perception information among autonomous vehicles and infrastructure, significantly enhancing perception capabilities by addressing limitations such as restricted range and occlusions inherent in on-board sensing. However, the limited bandwidth of V2X communication poses challenges for real-time transmission of the large data volumes required for early- and mid-level cooperative perception, necessitating consideration of communication efficiency alongside perception performance. To overcome this challenge, we propose a novel communication-efficient cooperative perception framework that achieves a better tradeoff between perception performance and communication cost. By exploiting the static deployment of infrastructure sensors and the heterogeneity of the perception region, the framework performs high-precision foreground segmentation on static infrastructure-side point clouds. A feature importance map is then constructed to extract sparse yet critical features for sharing, significantly reducing communication overhead. Additionally, we design a multiscale weighted attention feature fusion (MWAF) module to effectively integrate vehicle- and infrastructure-side features. The experimental results demonstrate that the proposed framework outperforms existing cooperative perception methods in terms of perception performance, while reducing communication traffic by approximately 60 times and 90 times on the DAIR-V2X and DZGSet dataset, respectively, without considering additional feature compression.

1 November 2026 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
LLLiwei Luo
Liwei Luo
XWXia WU
Xia WU