RT Journal Article
T1 Coronary artery segmentation based on transfer learning and UNet Architecture on Computed Tomography Coronary Angiography Images
A1 Serrano Antón, Belén
A1 Otero Cacho, Alberto
A1 López Otero, Diego
A1 Díaz Fernández, Brais
A1 Bastos-Fernández, María
A1 González Juanatey, José Ramón
A1 Pérez Muñuzuri, Vicente
K1 Artery
K1 Convolutional neural network
K1 Coronary
K1 CT
K1 Segmentation
K1 UNet
AB Coronary artery segmentation from CT scans is a helpful tool for coronary artery diseases diagnosis, which is frequently characterised by a vessel narrowing (stenosis). This is a highly demanded and high time-consuming process, thus automated procedures are becoming increasingly necessary. In this work, we propose an extremely light computationally 2D UNet that uses transfer learning for the first time in CT images. We compare the results, using different architectures and backbones, of a 2D UNet and a 3D UNet trained from scratch (i.e. weights are randomly initialised) and a 2D EfficientUNet. Both the amount of input data, with a total of 88 patients, and the extension of the structure to be recognised, the aorta and the coronary arteries ( A+C.A ), as well as the coronary arteries only ( C.A ) are analysed. Network outputs in clinically identified stenotic lesion areas are also assessed. The results show the advantage of using transfer learning when data is scarce, improving the F1 score by up to 0.6 points for the 2D UNet. On the other hand, when data is sufficient, F1 score values are close to 0.9 for all the networks. Besides, the results reveal that the 2D UNet distinguishes the thinnest and most distal vessels, although in the presence of a lesion, there is a clear tendency to overestimate it. The network with the best accuracy is the 3D UNet, with values above 95% and 75% in A+C.A and C.A , respectively. Moreover, the proposed methods show dependence on the amount of training data and dataset structure ( A+C.A or C.A ).
PB IEEE
YR 2023
FD 2023-07-26
LK https://hdl.handle.net/10347/37579
UL https://hdl.handle.net/10347/37579
LA eng
NO This work was supported in part by Spanish Ministerio de Economía y Competitividad and European Regional Development Fund under Contract RTI2018-097063-B-I00 AEI/FEDER, UE; in part by Xunta de Galicia under Grant 2021-PG036; in part by Spanish Ministerio de Ciencia e Innovación MCIN/AEI/10.13039/501100011033 through the Industrial Doctorates Grant and European Union NextGenerationEU/PRTR Research under Grant DIN2020-011068; and in part by FEDER (UE).
DS Minerva
RD 30 abr 2026