Javier Lazaro will defend his PhD Thesis entitled “ Development of computational tools to enhance Stimulated Raman histology for clinical applications” on Thursday, July 16, 2026 at 09:30 a.m. in Ponte Amphitheater, campus St Jérôme, Marseille.
Composition of the Jury :
Alexandre Kudlinski, Université de Lille – Jury President
Elsa Angelini, Telecom Paris – Reviewer
Charles Kervrann, INRIA – Reviewer
Senada Koljenovic, Antwerpen University Hospital – Examiner
Hervé Rigneault, CNRS Institut Fresnel – Supervisor
Julien Wojak, AMU Institut Fresnel – Co-supervisor
Rémi André, AMU Institut Fresnel – Co-supervisor
Abstract : Oncological diagnosis relies on microscopic analysis of tissue architecture, traditionally performed using hematoxylin and eosin (H&E) staining. Although effective, this approach is slow, labor-intensive, and destructive. Stimulated Raman Histology (SRH), based on Stimulated Raman Scattering microscopy, offers a rapid, label-free, and non-destructive alternative that highlights molecular contrasts associated with lipids and proteins, enabling virtual histology within minutes. However, its limited field of view, need for manual scanning, and unfamiliar image contrast compared to H&E constrain its clinical adoption. The work conducted in this PhD thesis aims to provide an automated framework for SRH image acquisition and interpretation. We introduce an intelligent microscopy strategy for automatic mosaicking of SRH field of views, guided by a U-Net-based model for cell nucleus segmentation. In addition, we develop a neural network for virtual staining to translate SRH images into H&E-like representations, facilitating pathological interpretation.
