Skip to Main content Skip to Navigation

Détection automatique des opacités en tomosynthèse numérique du sein

Abstract : Digital breast tomosynthesis is a new imaging technique that may potentially overcome some limitations of standard mammography like tissue superimposition. Unfortunately, the amount of data to review by the radiologist is also increased. In this context, it makes sense to design a tool to detect suspicious radiological findings in order to help him to have an acceptable reading time and to keep a high sensitivity. In this imaging modality, several patterns may indicate the presence of cancer: calcification clusters, masses and architectural distortions. In our work, we focus on the detection of masses and architectural distortions, which is challenging because of great variability and sometimes low contrast of these signs. In order to detect masses, we propose a sound theoretical work aiming at extending connected filters into the fuzzy sets framework. We also propose a completely different kind of tool in order to detect architectural distortions. This one is based on an a contrario modeling of the suspicious convergence detection problem. These tools are finally combined together in order to build a multi-channels detection system. The proposed approach is also compared to the state of the art techniques that can be found in the literature.
Complete list of metadata

Cited literature [257 references]  Display  Hide  Download
Contributor : Ecole Télécom ParisTech Connect in order to contact the contributor
Submitted on : Thursday, April 8, 2010 - 8:00:00 AM
Last modification on : Friday, October 23, 2020 - 4:37:50 PM
Long-term archiving on: : Thursday, March 30, 2017 - 5:45:00 AM


  • HAL Id : pastel-00005948, version 1



Giovanni Palma. Détection automatique des opacités en tomosynthèse numérique du sein. Traitement du signal et de l'image [eess.SP]. Télécom ParisTech, 2010. Français. ⟨pastel-00005948⟩



Record views


Files downloads