YMS Series #26 - Biomedical Raman-AI
Biomedical Raman-AI: From Spectral Preprocessing to Clinically Deployable Machine-Learning Workflows
Speaker: Bogdan Oancea
Date: 22 October 2026
Time: 18:00 Romanian time (EEST)
Abstract. Artificial intelligence is transforming biomedical research by enabling the extraction of clinically meaningful information from complex biological data. Among the emerging technologies benefiting from these advances, Raman spectroscopy offers a promising approach to cancer diagnostics through its label-free, non-destructive analysis of the molecular composition of biological tissues.
In this presentation, Prof. Bogdan Oancea will explore the complete computational pipeline connecting Raman spectral acquisition to clinically interpretable predictions. He will discuss the physical principles of Raman spectroscopy, spectral preprocessing and artifact removal, and the application of machine learning and deep learning to tissue classification, molecular characterization, and biomarker discovery.
Particular attention will be devoted to glioma classification and molecular subtyping, featuring the APOLLO framework, as well as RADAR, which illustrates how deep learning can improve spectral quality through advanced signal processing.
Beyond classification accuracy, the talk will address the challenges of translating experimental results into robust, reproducible, and clinically useful AI systems. Topics will include patient-level validation, data leakage, model interpretability, uncertainty quantification, synthetic data generation, and generalization across different acquisition conditions and instruments.
The central message is that successful clinical integration requires more than powerful algorithms. It demands carefully validated analytical workflows that preserve biological information and ensure scientific traceability from the original tissue specimen to the final prediction.
The presentation will highlight how the combination of spectroscopy, computational science, and artificial intelligence can contribute to more accurate, interpretable, and personalized approaches to cancer diagnosis.
Event organized by the Romanian Society of Bioinformatics.
Chair: Marius Surleac
Get notifications for other events: https://rsbi.ro/contact/
Speaker profiles:
ResearchGate: https://www.researchgate.net/profile/Bogdan-Oancea-2
Scopus: https://www.scopus.com/authid/detail.uri?authorId=26658614800
Web of Science: https://www.webofscience.com/wos/author/record/C-4147-2011
Google Scholar: https://scholar.google.com/citations?user=MHVWyc8AAAAJ
GitHub: bogdanoancea
#Bioinformatics #RSBI #ArtificialIntelligence #RamanSpectroscopy #MachineLearning #DeepLearning #BiomedicalAI #CancerDiagnostics