CV
Contact Information
| Name | Tong Wei |
| Professional Title | PhD candidate in Computer Vision |
| weitong@fel.cvut.cz |
Professional Summary
Computer vision PhD candidate working on learned matchability prediction for robust 3D reconstruction: selecting reliable image pairs in complex or noisy image collections for pose graph construction, SfM and feed-forward reconstruction, with related work on visual overlap prediction, global-aware edge scoring, feature matching and differentiable RANSAC.
Experience
-
2022 - present Prague, Czech Republic
Doctoral researcher
Visual Recognition Group, CTU in Prague
Learning-based methods for matchability prediction, reliable image-pair selection, pose graph construction and robust estimation for 3D reconstruction.
-
2026 - present Prague, Czech Republic
Researcher
SMART study, CTU in Prague
Machine-learning methods for oocyte quality and fertilization success prediction from polarized-light microscopy images.
-
2025 - 2025 Prague, Czech Republic
Researcher, sparse-to-fine billboard image verification
CTU in Prague
Sparse matching, homography estimation and dense patch verification for robust image verification.
-
2024 - 2024 Zurich, Switzerland
Academic guest
Computer Vision and Geometry Group, ETH Zurich
Minimum-spanning-tree-based image pair retrieval and pose graph construction for robust and efficient SfM.
- Supervised a semester project in the Mixed Reality course, “Where am I?” Interactive indoor localization and navigation on Magic Leap 2 hardware.
Education
-
2022 - present Prague, Czech Republic
PhD candidate
Czech Technical University in Prague
Computer Science
- Thesis: Learned Robust Matchability Prediction in Image Collections.
- Supervisors: Jiří Matas and Daniel Barath.
-
2020 - 2022 Budapest, Hungary
MSc
Eötvös Loránd University
Computer Science, AI specialization
- Grade: 4.36/5.
-
2015 - 2019 Liaoning, China
BSc
Liaoning Technical University
Communication Engineering
- Best Bachelor Thesis Award.
Publications
-
2026 Global-Aware Edge Prioritization for Pose Graph Initialization
CVPR 2026
T. Wei, G. Tolias, J. Matas, D. Barath. Oral presentation, Best Paper Award candidate (top 0.4% of submissions).
-
2025 Breaking the Frame: Visual Place Recognition by Overlap Prediction
WACV 2025
T. Wei, P. Lindenberger, J. Matas, D. Barath.
-
2023 Generalized Differentiable RANSAC
ICCV 2023
T. Wei, Y. Patel, A. Shekhovtsov, J. Matas, D. Barath.
-
2023 Adaptive Reordering Sampler with Neurally Guided MAGSAC
ICCV 2023
T. Wei, J. Matas, D. Barath.
-
2021 Wavelet Multi-level Attention Capsule Network for Texture Classification
IEEE Signal Processing Letters, 28:1215–1219
Z. Tao, T. Wei, J. Li.
-
2021 Underwater Image Sharpening Based on Structure Restoration and Texture Enhancement
Applied Optics, 60(15):4443–4454
S. Lin, K. Chi, T. Wei, Z. Tao.
Awards
-
2026 CVPR 2026 Best Paper Award candidate
CVPR
Global-Aware Edge Prioritization for Pose Graph Initialization.
-
2020 Stipendium Hungaricum Master's Scholarship
Hungarian Government
-
2021 Excellent Master's Student
Liaoning Technical University
-
2020 First Prize, Band A, National English Competition for College Students
China English Foreign Language Teachers Association
-
2019 Best Bachelor Thesis Award
Liaoning Technical University
3D Object Detection Network for Point Cloud Classification.
Talks
- Invited talk, Computer Vision Winter Workshop (CVWW), Slovenia, 2024: learning-based RANSAC scoring functions.
- Invited talk, Computer Vision Winter Workshop (CVWW), Austria, 2023: differentiable RANSAC components and applications.
Teaching and Service
- Teaching assistant, Medical Image Processing (BAM33ZMO), CTU in Prague, 2026–2027.
- Master thesis reviewer, CTU in Prague, 2025.
- Semester project supervisor, Mixed Reality course, ETH Zurich, 2024.
- Reviewer for CVPR, ECCV, NeurIPS, ICLR, AAAI and 3DV.
Summer Schools
- International Computer Vision Summer School (ICVSS), Sicily, Italy, 2023.
- Vision and Sports Summer School, Prague, Czech Republic, 2022.