Home > Actualite page d'accueil > BME Seminars - Elsa ANGELINI

BME Seminars - Elsa ANGELINI

Seminar announcement flyer

We are pleased to announce that our next seminar will be given by Elsa ANGELINI, lead Senior Data Scientist for the newly created Institute of Translational Medicine and Therapeutics (ITMAT) within NIHR Imperial Biomedical Research Centre (BRC). The seminar will be held between 1:30 pm and 3 pm in Amphi Bezier at ENSAM ParisTech this Friday, November 10, 2017. The seminar topic will be “Pulmonary emphysema quantification and subtyping on CT images: features and unsupervised learning on large cohorts”. For more information please refer to the attached abstract.

Pulmonary emphysema overlaps considerably with chronic obstructive pulmonary disease (COPD), and is traditionally subcategorized on CT scans into three subtypes: centrilobular emphysema (CLE), panlobular emphysema (PLE) and paraseptal emphysema (PSE). Automated classification methods based on supervised learning are generally based upon the current definition of emphysema subtypes, while unsupervised learning of texture patterns  enables the objective discovery of possible new radiological emphysema subtypes.
This talk will describe a 3-year adventure into unsupervised learning of emphysema subtypes and sub-phenotyping. We will go over image pixel information, visual features and building dictionaries, spatial sampling, use of clustering techniques and topics discovery. We will also review validation approaches when standard ground-truth is not available, evaluating reproducibility, prediction power and correlation with clinical measures.

Arts et Métiers ParisTech
Université Paris Descartes
Master BME Paris