Chengyang Zhang

Chengyang Zhang
Email: zhangchengy0705@gmail.com & Cy_Zhang@emails.bjut.edu.cn

About me

Hi! My name is Chengyang Zhang. I am a second-year Master's student at Beijing University of Technology, where I am grateful to be advised by Prof. Baocai Yin and Prof. Yong Zhang. During this period, I interned for half a year under the guidance of Prof. Bu Hong at West China Hospital, published one paper and obtained several patents.

Research Interests

I am broadly interested in the intersection area of Counting and Localization Technologies, Generative Models, and Graph Neural Networks. My current research focuses on medical image analysis and applications of generative models within the traffic domain. These include:

Cell Localization and Counting: Comprehensively iterated existing cell localization and counting paradigms, significantly advancing the progress of this task.

Traffic Generation with Diffusion Models: Designing high-performance diffusion models to generate traffic situations under the influence of abnormal events.

Selected Publications	(All publications see Google Scholar)
	Difference-Deformable Convolution with Pseudo Scale Instance Map for Cell Localization Chengyang Zhang, Jie Chen, Bo Li, Min Feng, Yongquan Yang, Qikui Zhu, Hong Bu Journal of Biomedical and Health Informatics, 2024 Code This paper introduces a novel gradient-aware and shape-adaptive Difference-Deformable Convolution, which enhances the model's robustness to color and substantial variability in cell morphology. Besides, the Pseudo-Scale Instance map is proposed, which adaptively provides the scale information of each cell for accurate supervision.
	BjTT: A Large-Scale Multimodal Dataset for Traffic Prediction Chengyang Zhang, Yong Zhang, Qitan Shao, Jiangtao Feng, Bo Li, Yisheng Lv, Xinglin Piao, Baocai Yin IEEE Transactions on Intelligent Transportation Systems, 2024 A large-scale multimodal dataset for traffic prediction. BjTT comprises over 32,000 time-series traffic records, capturing velocity and congestion levels on more than 1,200 roads within the 5th ring area of Beijing.
	CrowdGraph: Weakly supervised Crowd Counting via Pure Graph Neural Network Chengyang Zhang, Yong Zhang, Bo Li, Xinglin Piao, Baocai Yin ACM Transactions on Multimedia Computing, Communications, and Applications, 2024 This paper proposes a new graph-based crowd counting method named CrowdGraph to solve the problem of uneven distribution of crowd density, which reinterprets the weakly supervised crowd counting problem from a graph-to-count perspective.
	Multi-scale hypergraph-based feature alignment network for cell localization Bo Li, Yong Zhang, Chengyang Zhang, Xinglin Piao, Yongli Hu, Baocai Yin Journal of Pattern Recognition, 2024 Code This paper presents an innovative approach to address challenges arising from significant variations in cell shape, scale, and color. It reframes these challenges as a feature misalignment problem between cell images and location maps, offering a unified solution to these complexities.

Education

Beijing University of Technology (2022 - present)

Master in Electronic Information

Student Researcher at Beijing Institute of Artificial Intelligence

Awards: National Scholarship, The First Prize Scholarship

Beijing Information Science & Technology University (2018 - 2022)

Bachelor in Robotics Engineering

Awards: Excellent Graduate of Beijing; National Inspiration Scholarship; The First Prize Scholarship

Industry Research Experiences

West China Hospital, Chengdu, China (December 2022 - July 2023)

Algorithm Intern: Biomedical Image Analysis

Academic Service

Reviewer: IEEE Transactions on Intelligent Transportation Systems, Engineering Applications of Artificial Intelligence, Expert Systems With Applications.

Huge thanks for the website template from YueYANG1996.github.io.