Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

BP: Fachverband Biologische Physik

BP 14: Poster Session II

BP 14.58: Poster

Dienstag, 10. März 2026, 18:00–21:00, P2

Segmentation and classification of retinal pigment organelles in fluorescence lifetime imaging microscopy (FLIM) data — •Maryam Ali^{1, 2}, Hala Alhaj Ahmed^{3, 4}, Martin Hammer⁴, Rainer Heintzmann^{1, 2, 5}, and Ondrej Stranik^{1, 2} — ¹Leibniz Institute of Photonic Technology, Jena, Germany — ²Friedrich Schiller University, Jena, Germany — ³Ernst-Abbe University of Applied Sciences, Jena, Germany — ⁴Jena University Hospital, Jena, Germany — ⁵Abbe Centre of Photonics, Jena, Germany

Retinal Pigment Epithelium (RPE) granules can be categorized based on their autofluorescence and morphology like Lipofuscin (L), Melanolipofuscin (ML), and Melanin(M)[1]. Fluorescence lifetime measurements reveal another discriminative feature; however, identifying individual granules remain challenging by human eye. Here, we present a computational analysis pipeline for segmenting and classifying RPE granules from fluorescence lifetime imaging microscopy (FLIM) data. The analysis was implemented in a custom Python script employing seeded watershed segmentation to isolate individual granules and discriminate hyperfluorescent lipofuscins, characterized by longer lifetimes. Granules with shorter lifetimes were further analyzed by examining their lifetime distribution across their surfaces, allowing MLs to be distinguished from other melanin-rich granules. The proposed approach achieved high performance, with mean sensitivity 87% and mean specificity 98% compared to manually classified ground truth data. [1] K. Bermond et al., IOVS 2020, 61, 35

Keywords: Image segmentation; Image classification; Retinal pigment epithelium (RPE); Fluorescence lifetime imaging microscopy (FLIM); Autofluorescence