Ziheng Wang (王子恒)

I am currently an undergraduate advised by Xu Cheng (NUIST). I will soon begin my Ph.D. at the PCA Lab, advised by Jian Yang. My research focuses on building reliable visual intelligence systems, with interests spanning deep learning, computer vision, pattern recognition, and computer graphics. I am particularly passionate about connecting perception with real-world applications, and always open to academic collaboration. Feel free to reach out at zhwang@nuist.edu.cn.

Email  /  CV  /  Scholar  /  GitHub  /  Twitter

Research

My interests lie in intelligent perception, human-centered sensing, and reliable machine learning under real-world constraints such as noise and data imbalance. I enjoy transforming ideas into practical systems — from algorithms and datasets to deployable prototypes across software and embedded platforms.

We introduce FSCFNet on YOLOv10n with a novel frequency-spatial convolution (FSConv) using Haar wavelets. High-frequency branches enhance tiny-target details while low-frequency branches keep global context. An asymmetric cross-domain attention (ACA) module sharpens the typical Gaussian-like local patterns of IR small targets; customized multi-scale receptive designs improve speed-accuracy trade-offs.

We propose APPBoost, which introduces a parameter pair within a new loss and weight-update scheme to down-weight noisy/mislabeled samples while preserving hard informative cases. The method bounds training error, relates to Forward Stagewise Additive Modeling, and empirically improves generalization under noise/imbalance.

We propose AWABoost-SVM with two key parameters: one dynamically tunes the weight-update strength for harder samples; the other regularizes the SVM to balance complexity and error.

Projects

Health-tech ecosystem: remote vitals from camera, AI insights and coaching, and a modular UI.

PDF Introduction

Selected Patents

• Ziheng Wang, Zixuan Shao, Baowei Wang, Junqi Yang, Shiyu Wang, and Kaicen Tang. 一种用于心血管疾病诊断的双参数调节自适应增强方法. Chinese Invention Patent, Grant No. CN119128694B, Application No. 202411625421.3, Granted on Feb. 11, 2025. (Granted, First Inventor)

Selected Honors and Awards

• Jun 2026 — Outstanding Graduate of Nanjing University of Information Science & Technology
• Oct 2025 — National Scholarship for Undergraduates (国家奖学金, Top 2%)
• Oct 2025 — Three-Good Student Pacesetter, NUIST (Top 2%)
• Oct 2023 — National Second Prize, China Undergraduate Mathematical Contest in Modeling (Top 2%)
• May 2023 — First Prize, Jiangsu Advanced Mathematics Competition (Top 5%)

News

• Nov 2025 — Received awards from Prof. Haishan Chen, President of Nanjing University of Information Science & Technology, and Prof. Veronica Campbell, President of South East Technological University (Ireland) 🏅✨
• Nov 2025 — Honored as the university’s “Three-Good Student Pacesetter” (Top <2%) 🎉🎖️
• Oct 2025 — Awarded the National Scholarship for Undergraduates (Top 2%) 🎉🎊🎊
• Sep 2025 — Our paper was accepted by IEEE Transactions on Geoscience and Remote Sensing (TGRS, CCF-B) 🥳🥳✨
• Sep 2025 — Secured the No. 1 ranking in the school (Waterford, NUIST) for postgraduate recommendation quota (1/180, Top 0.55%), achieving a GPA of 4.0 and a comprehensive evaluation score of 106.12 🏆🥇
• Aug 2025 — Joined the PCA Lab as an incoming Ph.D. student 🎓
• Mar 2025 — Our APPBoost-based invention patent was officially granted 🎉🎊🎊
• Feb 2025 — Our paper was accepted by Journal of Supercomputing (CCF-C) 🥳🥳✨
• May 2024 — Led a project funded by the National Innovation & Entrepreneurship Training Program (Top 5%) 🎉🎊🎊
• Oct 2023 — Won the National Second Prize in the China Undergraduate Mathematical Contest in Modeling (Top 2%) 🎉🎊🎊
• May 2023 — Received the First Prize in the Jiangsu Advanced Mathematics Competition (Top 5%) 🎉🎊🎊