Fei Wu, Ph.D.

Lehrstuhl für Informatik 5 (Mustererkennung)

Research associates

Address

Martensstraße 3
91058 Erlangen
09.158  09

Fei Wu, Ph.D.

  • Since 09/2023:
    Postdoctoral researcher at the Pattern Recognition Lab, Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany.
  • 09/2017 – 07/2023:
    Doctor of Philosophy, School of Electrical, Electronics and Communication Engineering, University of Chinese Academy of Sciences, Beijing, China.
  • 09/2013 – 07/2017:
    Bachelor of Engineering, School of Computer Science and Engineering, Central South University, Changsha, China.

2023

  • Research on handwriting analysis, object tracking and segmentation based on machine learning

    (Own Funds)

    Project leader:
    Term: November 16, 2023 - May 1, 2026
    Acronym: b194dc-DocTraSeg

    This project investigates two computer vision tasks: 1. Handwriting analysis; 2. Object tracking and segmentation. 

    Handwriting analysis aims to evaluate and recognize the handwritten manuscripts according to different intentions, such as text recognition, spotting, layout analysis, text alignment, and writer recognition. As an important issue in the first step of digitizing scanned documents, this project will focus on layout analysis and line segmentation. 

    Object tracking and segmentation aims at continuously estimating the state of an object based on a given bounding box extracted by a simple rectangle/mask from the initial frame of a video sequence. It is widely applied in various applications such as surveillance, autonomous driving, human-computer interaction, etc. Despite the progress made so far, its main challenge lies in the limited discriminative power of the classifiers. Also, it is prone to the introduced endless distractors in real-world surveillance applications. For example, Siamese trackers dominate single-object-tracking field. Their balanced tracking paradigm coupled with fast inference speed and relatively high performance has caught the researchers’ attention. However, Siamese trackers mostly rely on large dataset offline training to learn the general representative capability for an arbitrary given target object. This ignores the target context relationship from adjacent frames. In addition, both CNNs and ViTs are used as feature extractor while the interaction between the local fine-grained and global coarse representation is still unexplored. This project will investigate state-of-the-art algorithms for achieving accurately and stably object tracking and segmentation.

2026

Journal Articles

2025

Journal Articles

Conference Contributions

2024

Journal Articles

Conference Contributions

2023

Journal Articles

2020

Journal Articles

2025

  • : Outstanding Reviewer of the 19th International Conference on Document Analysis and Recognition (ICDAR) – 2025
  • , : The Third Place Winner in ICDAR 2025 FEST Competition (ICDAR 2025 Competition on Few-Shot Text Line Segmentation of Ancient Handwritten Documents (FEST)) – 2025

2024

  • , : The First Place Winner in ICDAR 2024 SAM Competition (ICDAR 2024 Competition on Few-Shot and Many-Shot Layout Segmentation of Ancient Manuscripts (SAM)) – 2024

2023

  • : President’s Excellent Award of Chinese Academy of Sciences (CAS) – 2023

Current Theses & Projects

Title Type Student Period Status
Benchmarking 3D Transformers for Ink Detection in Carbonized Herculaneum Papyrus Scrolls MA thesis Muhammad Khalid Anwar Ahmed Omar Jun 2026 – Dec 2026 running
Frequency-Guided CNN-Transformer Hybridization for Handwritten Document Layout Analysis BA thesis Tasha Sanjay Khodanpur May 2026 – Oct 2026 running
Integrating Transformer Networks with Multi-Modal Learning for Document Layout Analysis MA thesis Yiqi Li Nov 2025 – Jun 2026 running

Completed Theses & Projects

Title Type Student Period Status
Benchmarking State-of-the-Art Transformers for Handwritten Document Layout Analysis MA thesis Nitesh Kumar Shah Sep 2025 – Apr 2026 finished