A novel scale-space framework for low-contrasted object segmentation: Empty catheter segmentation

Ketan Bacchuwar (1, 2) , Jean Cousty (2) , R Vaillant (1) , Laurent Najman (2)

1 GE Healthcare - General Electric Medical Systems [Buc]
2 LIGM - Laboratoire d'Informatique Gaspard-Monge

Purpose: In this article, we present empty guiding catheter segmentation in fluo-roscopic X-ray images. The guiding catheter, being a commonly visible landmark, its segmentation is an important block for Percutaneous Coronary Intervention (PCI) procedure modeling but difficult too. Methods: In number of clinical situations, it is empty and appears as a low contrasted structure with two parallel and partially disconnected edges. To segment it, we work on the level-set (non-linear) scale-space of image, the min tree, to extract curve blobs. We then propose a novel structural scale-space, a hierarchy built on these curve blobs. The deep connected component, i.e. the cluster of curve blobs on this hierarchy, that maximizes the likelihood to be an empty catheter is retained as final segmentation. Results: As a result of evaluation over a database of 1250 fluoroscopic images taken from examinations of 6 patients, we provide very good qualitative and quantitative segmentation performance, with mean precision and recall of 80.48% and 63.04% respectively. Conclusions: We develop a novel structural scale-space to segment a structured object, the empty catheter in challenging imaging situations where the information content is very sparse. Fully-automatic empty catheter segmentation in X-ray fluoroscopic images is an important and preliminary step in PCI procedure modeling, as it aids in tagging the arrival and removal of other interventional tools.

Domaines: Imagerie médicale; Traitement du signal et de l'image [eess.SP]; Imagerie

Ketan Bacchuwar, Jean Cousty, R Vaillant, Laurent Najman. A novel scale-space framework for low-contrasted object segmentation: Empty catheter segmentation. 2017. ⟨hal-01445849v1⟩