Wenyu Zhang

Mr. Wenyu Zhang

orcid

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 October 2023
Verified
Conference Paper
Open access
WZWenyu Zhang
Wenyu Zhang
XDXin Deng
Xin Deng
BJBaojun Jia
Baojun Jia
XYXingtong Yu
Xingtong Yu
YCYifan Chen
Yifan Chen
Xinming ZhangXinming Zhang
Xinming Zhang
Current Scene text image super-resolution approaches primarily focus on extracting robust features, acquiring text information, and complex training strategies to generate super-resolution images. However, the upsampling module, which is crucial in the process of converting low-resolution images to high-resolution ones, has received little attention in existing works. To address this issue, we propose the Pixel Adapter Module (PAM) based on graph attention to address pixel distortion caused by upsampling. The PAM effectively captures local structural information by allowing each pixel to interact with its neighbors and update features. Unlike previous graph attention mechanisms, our approach achieves 2-3 orders of magnitude improvement in efficiency and memory utilization by eliminating the dependency on sparse adjacency matrices and introducing a sliding window approach for efficient parallel computation. Additionally, we introduce the MLP-based Sequential Residual Block (MSRB) for robust feature extraction from text images, and a Local Contour Awareness loss (ℒlca) to enhance the model's perception of details. Comprehensive experiments on TextZoom demonstrate that our proposed method generates high-quality super-resolution images, surpassing existing methods in recognition accuracy. For single-stage and multi-stage strategies, we achieved improvements of 0.7% and 2.6%, respectively, increasing the performance from 52.6% and 53.7% to 53.3% and 56.3%. The code is available at https://github.com/wenyu1009/RTSRN.

15 December 2025 in Information
Verified
Journal Article
Open access
PDF file
XCXulei Cao
Xulei Cao
WZWenyu Zhang
Wenyu Zhang
FJFeiyang Jiang
Feiyang Jiang
Xinming ZhangXinming Zhang
Xinming Zhang
With the rise of online advertising, e-commerce industries, and new media platforms, recommendation systems have become an essential product form that connects users with a vast number of candidates. A major challenge in recommendation systems is the cold-start problem, where the absence of historical interaction data for new users and items leads to poor recommendation performance. We first analyze the causes of the cold-start problem, highlighting the limitations of existing embedding models when faced with a lack of interaction data. To address this, we classify the features of models into three categories, leveraging the Trans Block mapping to transfer features into the semantic space of missing features. Then, we propose a model-agnostic industrial framework (MAIF) with the Auto-Selection serving mechanism to address the cold-start recommendation problem in few-shot and zero-shot scenarios without requiring training from scratch. This framework can be applied to various online models without altering the prediction for warm entities, effectively avoiding the “seesaw phenomenon” between cold and warm entities. It improves prediction accuracy and calibration performance in three cold-start scenarios of recommendation systems. Finally, both the offline experiments on real-world industrial datasets and the online advertising system on the Dazhong Dianping app validate the effectiveness of our approach, showing significant improvements in recommendation performance for cold-start scenarios.

21 February 2026
Verified
Conference Paper
Open access
JLJie Luo
Jie Luo
WZWenyu Zhang
Wenyu Zhang
Xinming ZhangXinming Zhang
Xinming Zhang
YFYuan Fang
Yuan Fang

16 December 2024 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
GYGuoliang You
Guoliang You
XCXiaomeng Chu
Xiaomeng Chu
YDYifan Duan
Yifan Duan
WZWenyu Zhang
Wenyu Zhang
XLXingchen Li
Xingchen Li
YZYanyong Zhang
Yanyong Zhang
When planning for autonomous driving, it is crucial to consider essential traffic elements such as lanes, intersections, traffic regulations, and dynamic agents. However, they are often overlooked by the traditional end-to-end planning methods, likely leading to inefficiencies and non-compliance with traffic regulations. In this work, we endeavor to integrate the perception of these elements into the planning task. To this end, we propose P erception H elps P lanning (PHP), a novel framework that reconciles lane-level planning with perception. This integration ensures that planning is inherently aligned with traffic constraints, facilitating safe and efficient driving. Specifically, PHP focuses on both edges of a lane for planning and perception purposes, taking into account the positions of both lane edges in Bird's Eye View (BEV), along with attributes related to lane intersections, lane directions, and lane occupancy. In the algorithmic design, the process begins with the transformer encoding multi-camera images to extract the above features and predict lane-level perception results. Next, the hierarchical feature early fusion module refines the features for predicting planning attributes. Finally, a specific interpreter utilizes a late-fusion process designed to integrate lane-level perception and planning information, culminating in generating vehicle control signals. Experiments on three Carla benchmarks show significant improvements in driving scores of 27.20%, 33.47%, and 15.54% over existing algorithms, respectively, achieving state-of-the-art performance, with the system operating up to 22.57 FPS.


16 September 2023 in arXiv
Verified
Preprint
Open access
PDF file
WZWenyu Zhang
Wenyu Zhang
XDXin Deng
Xin Deng
BJBaojun Jia
Baojun Jia
XYXingtong Yu
Xingtong Yu
YCYifan Chen
Yifan Chen
Xinming ZhangXinming Zhang
Xinming Zhang
Current Scene text image super-resolution approaches primarily focus on extracting robust features, acquiring text information, and complex training strategies to generate super-resolution images. However, the upsampling module, which is crucial in the process of converting low-resolution images to high-resolution ones, has received little attention in existing works. To address this issue, we propose the Pixel Adapter Module (PAM) based on graph attention to address pixel distortion caused by upsampling. The PAM effectively captures local structural information by allowing each pixel to interact with its neighbors and update features. Unlike previous graph attention mechanisms, our approach achieves 2-3 orders of magnitude improvement in efficiency and memory utilization by eliminating the dependency on sparse adjacency matrices and introducing a sliding window approach for efficient parallel computation. Additionally, we introduce the MLP-based Sequential Residual Block (MSRB) for robust feature extraction from text images, and a Local Contour Awareness loss ($\mathcal{L}_{lca}$) to enhance the model's perception of details. Comprehensive experiments on TextZoom demonstrate that our proposed method generates high-quality super-resolution images, surpassing existing methods in recognition accuracy. For single-stage and multi-stage strategies, we achieved improvements of 0.7\% and 2.6\%, respectively, increasing the performance from 52.6\% and 53.7\% to 53.3\% and 56.3\%. The code is available at https://github.com/wenyu1009/RTSRN.