I'm LukΓ‘Ε‘ Picek, a computer vision and applied AI researcher and PI at the University of West Bohemia π¨πΏ.
My research sits at the intersection of computer vision, machine learning, and biodiversity science. I develop methods, datasets, benchmarks, and open-source tools to help us better understand the natural world, from identifying individual animals and recognizing species to predicting where they occur.
Over the past three years, I've had the opportunity to work with some great people at Inria π«π·, as part of the Pl@ntNet team, and MIT CSAIL πΊπΈ, where I was a Fulbright Visiting Scholar in the Beery Lab.
I particularly enjoy collaborating with ecologists, biologists, and other domain experts to tackle scientific challenges where AI can make a difference. I'm also interested in what happens beyond the benchmark: whether our models generalize, how their errors affect scientific conclusions, and how to turn research into tools that others can actually use.
- π§ͺ Trustworthy Benchmarks & Evaluation β Developing datasets and evaluation protocols that reveal model limitations, assess generalization, and measure performance beyond conventions.
- πΎ Animal re-identification & population monitoring β Recognizing individual animals and connecting AI with ecological research.
- π Fine-grained species recognition β Identifying plants, fungi, and animals in challenging long-tailed and open-world settings.
- π Species distribution modeling β Combining biodiversity data with environmental and spatial data to predict species distributions.
A few projects I've developed or contributed to:
| Project | Description |
|---|---|
| MegaDescriptor | Foundation models for animal re-identification across species. |
| WildlifeDatasets | Open-source toolkit for animal re-identification datasets, evaluation, and benchmarking. |
| WildlifeTools | Models and practical tools for individual animal identification. |
| Malpolon | Deep learning framework for species distribution modeling. |
| FungiTastic | Large-scale dataset and benchmark for fine-grained fungi recognition. |
| GeoPl@ntNet | Platform for exploring plant distributions, biodiversity indicators, and ecological data. |
I enjoy bringing researchers together and helping build communities around challenging problems.
I'm the lead organizer of **LifeCLEF**and the Fine-Grained Visual Categorization (FGVC) workshop at CVPR, connecting people working on computer vision, machine learning, and biodiversity.
I'm also interested in developing better ways to evaluate AI for scientific research (e.g. Terminal Bench Science), including benchmarks that go beyond conventional machine learning metrics.
Research website Β· Google Scholar Β· LinkedIn Β· Kaggle
Always happy to connect, collaborate, or exchange ideas about computer vision, biodiversity, and AI for science.