From Microscopy to Medicine: AI Approaches for Biomedical Image Analysis

Apstrakt: 

This presentation explores the growing role of signal processing, machine learning, and artificial intelligence in biomedical data analysis and medical imaging. The first part focuses on computational methods for microscopy-based biological research, including multiview deconvolution for image reconstruction, automated cell segmentation and tracking, and phenotyping approaches based on cellular morphology. In addition, matrix factorization combined with active learning is presented as an efficient strategy for drug target prediction and discovery from high-dimensional biological data.

The second part addresses challenges in digital pathology, with a focus on deep learning methods for histopathological image segmentation. In particular, approaches for glomeruli detection in kidney tissue images are discussed, alongside the use of Generative Adversarial Networks (GANs) for stain normalization and domain adaptation to improve segmentation performance across different staining protocols and datasets.

Finally, the presentation examines the broader impact of artificial intelligence in medicine, highlighting its emerging role as a diagnostic support tool and decision-making assistant. The talk concludes with a discussion on the future of AI and the potential development of Artificial General Intelligence (AGI) systems for medicine, emphasizing opportunities, limitations, and ethical considerations in next-generation healthcare technologies.