Feasibility study of catheter segmentation in 3D Frustum Ultrasounds by DCNN

Lan Min, Hongxu Yang, Caifeng Shan, Alexander F. Kolen, Peter de With

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

Samenvatting

3D ultrasound has been developed rapidly in medical intervention therapies, such as cardiac catheterization. Image-based catheter detection is studied to help sonographer to timely localize the instrument in the 3D US images. Conventionally, the 3D imaging methods are based on the Cartesian domain, which is limited by bandwidth and information lose when it is converted from the original acquisition space---Frustum domain. The catheter segmentation in the Frustum space helps to reduce the computational cost and improve efficiency. In this paper, we present a catheter segmentation method in 3D Frustum image via a deep convolutional network (DCNN). To accelerate the prediction efficiency on whole US Frustum volume, a filter-based pre-selection is applied to reduce the computational cost of the DCNN. Based on experiments on the ex-vivo dataset, our proposed method can segment the catheter in Frustum images with 0.67 Dice score within 3 seconds.
Originele taal-2Engels
TitelImage-Guided Procedures, Robotic Interventions, and Modeling
UitgeverijSPIE
Aantal pagina's6
DOI's
StatusGepubliceerd - 16 mrt 2020
Evenement2020 SPIE Medical Imaging: Image-Guided Procedures, Robotic Interventions, and Modeling - Houston, Verenigde Staten van Amerika
Duur: 16 feb 202019 feb 2020

Publicatie series

NaamProceedings of SPIE
Nummer11315

Congres

Congres2020 SPIE Medical Imaging
LandVerenigde Staten van Amerika
StadHouston
Periode16/02/2019/02/20

    Vingerafdruk

Citeer dit

Min, L., Yang, H., Shan, C., Kolen, A. F., & de With, P. (2020). Feasibility study of catheter segmentation in 3D Frustum Ultrasounds by DCNN. In Image-Guided Procedures, Robotic Interventions, and Modeling [1131521] (Proceedings of SPIE; Nr. 11315). SPIE. https://doi.org/10.1117/12.2549084