1. Multi-modal Auto-labelling Pipeline for Sugar Beet Crop-Weed Classification

The deployment of robust machine learning models in precision agriculture is frequently bottlenecked by the scarcity of high-quality annotated data. Despite recent developments in ML-based crop–weed detection systems, the availability of large-scale, high-quality labelled datasets remains a major limitation. State-of-the-art models require extensive data that captures diverse growth stages, variable lighting and weather conditions, and a wide range of weed species. Creating such datasets... A. Zakhary, S. Henkler, A. Ehteshami-bejnordi, D. Yablonski, V. Sorokina