I am currently a fourth-year Ph.D. student (2022 – now) in the Department of Automation, University of Science and Technology of China (USTC), advised by Prof. Tianzhu Zhang and Prof. Baoqun Yin. I spent my M.Eng. studies at the Advanced Technology Research Institute, USTC. Huazhong University of Science and Technology conferred upon me a B.Eng. degree in 2020 from the School of Electrical and Electronic Engineering. My research interests include computer vision, multimodal learning and 3D vision.

🔥 News

📝 Publications

GLASS: Geometry-aware Local Alignment and Structure Synchronization Network for 2D-3D Registration Zhixin Cheng, Jiacheng Deng, Xinjun Li, Bohao Liao, Li Liu, Xiaotian Yin, Baoqun Yin, Tianzhu Zhang* This paper revisits image-to-point cloud registration by addressing two key challenges: low-texture or repetitive regions causing mismatches, and inherent modality differences between images and point clouds. To tackle these, GLASS introduces a Local Geometry Enhancement module that injects surface normals to enhance structural awareness. It further applies a Graph Distribution Consistency module to regularize the similarity distributions of matched keypoints. By combining local geometric enhancement with structural consistency, GLASS reduces ambiguity and incorrect matches. Presented at IEEE Transactions on Circuits and Systems for Video Technology🌸

Rethinking 2D-3D Registration: A Novel Network for High-Value Zone Selection and Representation Consistency Alignment Zhixin Cheng, Bohao Liao, Jiacheng Deng, Xiaotian Yin, Xinjun Li, Yujia Chen, Baoqun Yin, Tianzhu Zhang* This paper rethinks image-to-point cloud registration by addressing two key challenges: many correspondences arise from non-overlapping or low-quality regions, and the same scene areas appear differently across modalities since images capture texture while point clouds represent geometry. The method therefore emphasizes selecting informative, matchable regions first, and then enforcing more consistent cross-modal region representations, reducing ambiguity and mismatches. Presented at Computer Vision and Pattern Recognition 2026🌸

Adaptive Agent Selection and Interaction Network for Image-to-Point Cloud Registration Zhixin Cheng, Xiaotian Yin, Jiacheng Deng, Bohao Liao, Yujia Chen, Xu Zhou, Wenfei Yang*, Baoqun Yin This paper targets the challenges of image-to-point-cloud registration under noise, where false correspondences are common and cross-modal information is difficult to filter effectively. It proposes a framework composed of Iterative Agent Selection and Reliable Agent Interaction: phase maps enhance structural perception, and reinforcement learning selects more reliable agents to guide cross-modal interaction, thereby reducing mismatches and improving robustness. Presented at Association for the Advancement of Artificial Intelligence 2026🌸

CA-I2P: Channel-Adaptive Registration Network with Global Optimal Selection Zhixin Cheng, Jiacheng Deng, Xinjun Li, Xiaotian Yin, Bohao Liao, Baoqun Yin, Wenfei Yang*, Tianzhu Zhang This paper addresses detection-free image-to-point cloud registration, where cross-modal channel mismatches and redundant top-k correspondences reduce matching quality. CA-I2P uses a Channel Adaptive Adjustment module to align channels across modalities and a Global Optimal Selection module to produce globally consistent matches for robust registration. We attend the conference and discussed our ideas with Google AI scientist Martin Sundermeyer. Presented at International Conference on Computer Vision 2025🌸

Bridge 2D-3D: Uncertainty-aware Hierarchical Registration Network with Domain Alignment Zhixin Cheng, Jiacheng Deng, Xinjun Li, Baoqun Yin, Tianzhu Zhang* This paper proposes B2-3Dnet for detection-free image-to-point cloud registration, aiming to reduce distraction from noisy image patches and narrow the cross-modal domain gap. It introduces an uncertainty-aware hierarchical matching module that estimates patch reliability and performs multi-scale coarse-to-fine interactions, and an adversarial modal alignment module that aligns image and point-cloud features using a gradient reversal strategy and a domain classifier. Presented at Association for the Advancement of Artificial Intelligence 2025🌸

GeoGuide: Hierarchical Geometric Guidance for Open-Vocabulary 3D Semantic Segmentation

Xujing Tao, Chuxin Wang, Yubo Ai, Zhixin Cheng, Zhuoyuan Li, Liangsheng Liu, Yujia Chen, Xinjun Li, Qiao Li, Wenfei Yang*, Tianzhu Zhang

RayI2P: Learning Rays for Image-to-Point Cloud Registration

Xinjun Li, Wenfei Yang*, Zhixin Cheng, Jiacheng Deng, Fei Wang, Chen Qian, Tianzhu Zhang

Adversarial Attacks Already Tell the Answer: Directional Bias-Guided Test-time Defense for Vision-Language Models

Liangsheng Liu, Si Chen, Jiamin Wu, Weiwei Feng, Zhixin Cheng, Xiaotian Yin, Wenfei Yang*, Tianzhu Zhang

EF-3DGS: Event-Aided Free-Trajectory 3D Gaussian Splatting (Spotlight)

Bohao Liao, Wei Zhai*, Zengyu Wan, Zhixin Cheng, Wenfei Yang, Yang Cao, Tianzhu Zhang, Zhengjun Zha

BeyondMix: Leveraging Structural Priors and Long-Range Dependencies for Domain-Invariant LiDAR Segmentation

Yujia Chen, Rui Sun, Wangkai Li, Huayu Mai, Si Chen, Zhuoyuan Li, Zhixin Cheng, Tianzhu Zhang*

Implicit Correspondence Learning for Image-to-Point Cloud Registration (Highlight)

Xinjun Li, Wenfei Yang*, Jiacheng Deng, Zhixin Cheng, Xu Zhou, Tianzhu Zhang

DiffCorr: Conditional Diffusion Model with Reliable PseudoLabel Guidance for Unsupervised Point Cloud Shape Correspondence

Jiacheng Deng, Jiahao Lu, Zhixin Cheng, Wenfei Yang*

🏅 Honors and Awards

Outstanding Graduate of Anhui Province Ordinary Higher Education Institutions, Class of 2026

Outstanding Graduate of the Class of 2026, USTC

First‑Class Scholarship, USTC Graduate School

Deep Space Exploration Scholarship

Excellent Minister, Graduate Student Union, Advanced Technology Research Institute, USTC

USTC “复兴杯” Basketball Champion (2024) and Runner‑up (2023)

Outstanding Graduate, Huazhong University of Science and Technology

Level 10 Certification in Erhu Performance🎶

📖 Education

2022.09 – present: Ph.D. in Automation, USTC, Hefei, China

2020.09 – 2022.08: M.Eng., Advanced Technology Research Institute, USTC, Hefei, China

2016.09 – 2020.06: B.Eng., School of Electrical and Electronic Engineering, Huazhong University of Science and Technology, Wuhan, China

2013.09 – 2016.06: Senior High School, Hefei No.8 High School, Hefei, China

2010.09 – 2013.06: Junior High School, Hefei No.50 Middle School, Hefei, China

💻 Internships

2025.06 – 2025.10: Research Intern, COG1, Spark Large Model Research Institute, iFLYTEK, Hefei, China

2021.09 – 2022.08: Research Intern, Brain‑Inspired Intelligence Platform, Hefei Comprehensive National Science Center, Hefei, China

2020.09 – 2021.03: AI Algorithm Intern, Jiyuan–Huawei Ascend Joint Laboratory, NARI Jiyuan Electric Grid Technology Co., Ltd., Hefei, China

2018.07 – 2018.09: Visiting Student, The University of Manchester, Manchester, UK🎡

🛠 Skills

Programming: Python, PyTorch, CUDA

Research: Deep learning, Multimodal fusion, 3D vision task

Service: Reviewer for ICML, CVPR, ICCV, AAAI, NIPS, ICLR, ACM MM, TCSVT, TMM

Language: CET‑6, good English writing and communication

📬 Contact

Email: chengzhixin@mail.ustc.edu.cn