Fei Cheng (程飞)

Assistant Professor · Dept. of Communication and Networking · Xi’an Jiaotong-Liverpool University (XJTLU), Suzhou, China

Computer Vision SLAM Spatial Computing Deep Learning Image/Video Processing
Email
Office
Room 216, No. 111 Renai Road, Suzhou, Jiangsu, China
Phone
+86 135 8443 5676

About

academic profile

I am an Assistant Professor at Xi’an Jiaotong-Liverpool University (XJTLU), Department of Communication and Networking. My work spans computer vision, SLAM, and learning-based spatial understanding, with applications in intelligent construction and industrial inspection.

Research Interests

focus areas
  • Visual SLAM & panoramic (equirectangular) camera geometry
  • 3D reconstruction / 3D Gaussian Splatting and spatial computing
  • Deep neural networks for detection, tracking, and quality assessment
  • Image/video coding, motion modeling, and perception-inspired processing

Teaching

modules
  • CAN201: Introduction to Networking
  • CPT404: Technological Project Management

Projects

selected
  • 3D Gaussian Splatting Model Based Object Segmentation and Intelligent Measurement for Intelligent Construction (Industry Project, PI, 2025–2027)
  • Pathology Slice Analysis Software Based on Deep Learning Models (Industry Project, PI, 2025–2026)
  • Low-code framework R&D for research affairs management (SAT fund, PI, 2022–2024)

Selected Publications

highlights
  • DFF: Sequential Dual-Branch Feature Fusion for Maritime Radar Object Detection and Tracking via Video Processing. Applied Sciences, 2025.
  • A Parallel Corpus of Chinese-English Legal Judgments with Argumentative Structure Annotations. Data Intelligence, 2025.
  • A Content-Aware Full-Reference Image Quality Assessment Method Using a Gram Matrix and Signal-to-Noise. IEEE Transactions on Broadcasting, 2024.
  • Texture plus Depth Video Coding Using Camera Global Motion Information. IEEE Transactions on Multimedia, 2017.

For the full list, please refer to FeiCheng_CV_2022.pdf.

Patents

selected
  • An exam paper marking aggregation system and method based on random QR code. CN202010417727.5 (Granted, 2024).
  • Construction site elevator population counting system and method based on deep neural network. CN202110243399.6 (Granted, 2021).

Contact

reach me