CV

Klára Janoušková

Prague, Czech Republic · klara.janou@gmail.com · LinkedIn · GitHub · Google Scholar · Bluesky

Download PDF

Summary

Computer vision researcher with experience in both academia and industry. I currently work on 3D instance segmentation and robotics, vision-language model (VLM) adaptation, trustworthy benchmarks, and RL for visual object tracking, and have co-led several research projects from the initial idea to publication.

Education

Ph.D. in Computer Vision & Machine Learning — Czech Technical University in Prague
2023 – present
  • Supervisor: Prof. Jiří Matas
  • Passed state doctoral exams with distinction (top 10%)
M.Sc. in Computer Science, graduated with honors — Czech Technical University in Prague
2021 – 2023
  • Thesis: Test-Time Adaptation for Segmentation
B.Sc. in Open Informatics — Czech Technical University in Prague (CTU)
2017 – 2021
Prague, Czech Republic
  • Thesis: Scene Text Detection and Recognition — link · Dean's Award

Experience

Senior Researcher — Equilibre Labs @ Equilibre Technologies
Oct 2026 – present
Prague, Czech Republic
  • Supervise student projects at the intersection of computer vision and robotics.
PhD Researcher — Czech Technical University in Prague
2023 – present
  • Mentoring BSc/MSc/high-school students; Teaching Assistant for ML & Pattern Recognition.
  • Secured a Toyota Motors Europe Research Grant for a one-year project on efficient vision-language models.
Research Intern — Technion (Israel Institute of Technology)
Sep – Oct 2022
Haifa, Israel
  • Selected via the Czech-Israeli Innovation Internship (CI3) program, hosted by Alex Bronstein and Chaim Baskin.
  • Worked on single-image test-time adaptation for segmentation; published in TMLR 2024.
Research Intern — Equilibre Technologies
2022
Prague, Czech Republic
  • Collaboration with an algorithmic trading company on ML research during M.Sc.
Research Intern — IBM Research Zurich
July – Dec 2021
Zürich, Switzerland
  • Selected via IBM Great Minds fellowship, worked as part of the AI automation group led by Cristiano Malossi.
  • Best Paper Award, ECCV 2022 Workshop on Computer Vision for Civil and Infrastructure Engineering.
Research Intern — Computer Vision Center (CVC) at UAB
July – Sep 2019
Barcelona, Spain
  • Hosted by Dimosthenis Karatzas.
  • Worked on scene-text detection and recognition; published at ICPR 2020.

Teaching

Teaching Assistant — Czech Technical University in Prague
2023 – present
Prague, Czech Republic
  • Labs for the Machine Learning and Pattern Recognition course.
  • Co-supervision of BSc, MSc, and high-school students.

Selected Publications

† corresponding author / project lead. Full list on Google Scholar.

  1. Klara Janouskova, Jiri Matas, Federico Tombari, Keisuke Tateno, Daniel Barath. “Tomos3D: Zero-Shot 3D Instance Segmentation from Cutting Plane Renderings.” Under review (ICLR), 2027.
  2. Klara Janouskova, Zuzana Moravčíková, Nikita Kisel, Jiri Matas. “SEAR-VLM: Semantic Embedding for Any Region Learned with Structured MLLM Annotations.” Under review (AAAI), 2027. [dataset]
  3. Tomáš Čížek, Matej Straka, Illia Volkov, Matěj Kripner, Ervin Macić, Jiri Matas, Klara Janouskova, Martin Schmid. “SAM3RL: Memory Control via Reinforcement Learning for Visual Object Tracking.” Under review (ICLR), 2027. [link]
  4. Illia Volkov, Nikita Kisel, Tetyana Mishkina, Klara Janouskova†, Jiri Matas. “Doomed to Re-Annotate, Forever: The ImageNet Story.” NeurIPS Evaluations & Datasets Track (Spotlight), 2026.
  5. Nikita Kisel, Illia Volkov, Klara Janouskova†, Jiri Matas. “Multimodal LLMs as Image Classifiers.” CVPR Findings, 2026. [link]
  6. Nikita Kisel, Illia Volkov, Katerina Hanzelkova, Klara Janouskova†, Jiri Matas. “Flaws of ImageNet, Computer Vision's Favourite Dataset.” ICLR Blog Posts, 2025. [link]
  7. Klara Janouskova, Jiri Matas, Lukas Picek. “Overview of FungiCLEF 2025: Few-Shot Classification with Rare Fungi Species.” Working Notes of CLEF, 2025.
  8. Lukas Picek, Klara Janouskova, Vojtech Cermak, Jiri Matas. “FungiTastic: A Multi-modal Dataset and Benchmark for Image Categorization.” CVPR workshops, Fine-Grained Visual Categorization, 2025. [link]
  9. Klara Janouskova, Tamir Shor, Chaim Baskin, Jiri Matas. “Single Image Test-Time Adaptation for Segmentation.” TMLR, 2024. [link]
  10. Klara Janouskova, Mattia Rigotti, Ioana Giurgiu, Cristiano Malossi. “Model-Assisted Labeling via Explainability for Visual Inspection of Civil Infrastructures.” ECCV Workshops, Computer Vision for Civil and Infrastructure Engineering (Best Paper Award), 2022. [link]
  11. Klara Janouskova, Jiri Matas, Lluis Gomez, Dimosthenis Karatzas. “Text Recognition: Real World Data and Where to Find Them.” ICPR, 2020. [link]

Patents

  • Klára Janoušková, Ioana Giurgiu, Mattia Rigotti, Cristiano Malossi. Generating Strong Labels for Examples Labelled with Weak Labels. US Patent 12,406,478. Granted 2025.

Skills

Languages
Python, Bash, LaTeX
Frameworks
PyTorch, Hugging Face Transformers, OpenCV
Areas
Computer Vision, Machine Learning, Zero-Shot Segmentation, Test-Time Adaptation, Fine-Grained Classification

Awards & Honours

  • Outstanding Reviewer, CVPR 2026
  • Top Reviewer, NeurIPS 2024
  • Josef Hlavka Award, The Foundation of Josef, Marie and Zdeňka Hlávka, 2022
    Awarded to outstanding students and graduates (oldest Czech foundation supporting science and culture).
  • Best Paper Award, ECCV 2022 Workshop on Computer Vision for Civil and Infrastructure Engineering
  • Dean's Award for Outstanding Bachelor's Thesis, Czech Technical University, 2020
  • 100th Percentile, SCIO National Comparative Exams in Mathematics, 2017
    Best result in the cohort; used as an admission criterion at many Czech universities.
  • Open Informatics Scholarship for Talented Girls, 2017
    Awarded to top-scoring students in the NCE in Mathematics.

Grants & Funding

  • Toyota Motors Europe Research Grant, €100,000 (2025–2026)
    One-year research project on efficient vision-language models.

Professional Service

Organizing
Workshop on Fine-Grained Visual Categorization (CVPR 2025), FungiCLEF Challenge 2025
Reviewing
CVPR (Outstanding Reviewer 2026), NeurIPS (Top Reviewer 2024), ECCV, ICCV, IJCV, WACV, PUT@ICML, FGVC@CVPR