Bionote
Dr. Eva Maria Molin, MSc, born 29 December 1980 in Klagenfurt, AT
1999-2006 studied biology at the University of Vienna. 2006 Master’s degree in Biology, 2011 PhD at the Gregor Mendel Institute for Molecular Plant Sciences, Vienna.
Since 2013 scientist and project manager at the AIT Austrian Institute of Technology, Center for Health & Bioresources, Unit Bioresources. 2015-2016 interim head of the Business Case “DNA Bank & Genotyping Services”. 2018-2020 part-time studies at the FH Campus Wien. 2020 MSc in Bioinformatics. Since 2024 Member of the AIT AI Taskforce and Visiting Researcher at the Human-Centered AI Lab Vienna, BOKU University. Certified mountain hiking guide since 2025.
Presentation Abstract
AIT Austrian Institute of Technology, Center for Health & Bioresources, Unit Bioresources
Wheat is one of the world’s most important staple crops, but increasingly unfavourable environmental conditions linked to climate change pose a serious threat to yield stability and food security. To address this challenge, the WheatVIZ1 project focuses on accelerating wheat breeding processes by integrating advanced UAV-based phenotyping methods with genomic analysis and explainable machine learning (ML) techniques. The ultimate goal is to develop user-friendly, resource-efficient methods and tools that can be seamlessly integrated into breeding programmes to increase their effectiveness in selecting stress-tolerant wheat genotypes.
In this process, phenotyping is an important step, and as such, UAV-based phenotyping is being increasingly applied. It enables scalable, non-destructive phenotyping of various agronomic, morphological and physiological traits, significantly reducing time and labour compared to traditional methods. By implementing a novel deep learning architecture—TriNET2—including interpretable ML models, the project optimises phenotyping strategies for yellow leaf rust through selective spectral bands and efficient flight planning, achieving robust predictive performance and actionable insights for agronomists and plant breeders. This scalable and explainable approach provides an efficient solution and has the potential to be adapted for other phenotypes and species.
Co-Authors: Ignacio Chang-Brahim, Lukas J. Koppensteiner, Lorenzo Beltrame, Gernot Bodner, Anna Saranti, Jules Salzinger, Philipp Fanta-Jende, Christoph Sulzbachner, Felix Bruckmüller, Friederike Trognitz, Mina Samad-Zamini, Elisabeth Zechner, Andreas Holzinger