Valerio Giuffrida is an Assistant Professor in Computer Vision, recently appointed to the School of Computer Science at the University of Nottingham. His academic activities have mostly focused on leveraging computer vision and deep learning methodologies to address challenges in plant image analysis. All started by proposing a machine learning-based approach for leaf counting in images of Arabidopsis thaliana plants, establishing himself as a pioneer in the field.

Throughout his career, Valerio has expanded been involved in other related research projects, including root phenotyping, multimodal learning (not only on plant image analysis, but also in medical imaging), generative models, and foundation models. He currently a Co-Investigator for the PhenomUK Scoping Project, funded by BBSRC, where he is involved in the Data Infrastructure strand alongside Sotos Tsaftaris from the University of Edinburgh.

Valerio has actively contributed to the organisation of several international events. Currently, he is organising the 9th edition of the Computer Vision Problems in Phenotyping and Agriculture Workshop, which could happen at ECCV ’24 (subjected to approval).

---

About this seminar

 In his presentation, Valerio discusses the challenges associated with training machine learning algorithms for leaf counting tasks.

Throughout this talk, he shows several methodologies, ranging from regression to segmentation, showcasing how these techniques enable machine learning algorithms to accurately count leaves from plant images. Starting with traditional machine learning approaches, Valerio explores the potential of advanced models such as transformers.

By leveraging these techniques, more robust algorithms capable of tackling complex counting tasks can be developed across diverse domains. From plant phenotyping to applications in biology and CCTV systems.