http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Enlarged perivascular spaces segmentation in brain MRI with a deep neural network
Shin Won Kang(강신원),Ehwa Yang(양이화),Won-Jin Moon(문원진),Yeonsil Moon(문연실),Hee-Jin Kim(김희진),Jae-Hun Kim(김재훈) 한국통신학회 2022 한국통신학회 학술대회논문집 Vol.2022 No.2
Enlarged perivascular space (EPVS) is one of the biomarkers of vascular brain diseases. Quantitative study of EPVS in the brain is important for understanding its associations with the disease, and for clinical diagnosis. Due to their tiny and sparse appearances together with their shared intensities with other types of brain lesions on MRI, automatic methods to detect EPVSs are strongly challenged. In this study, we employ a residual U-Net for the segmentation of basal ganglia EPVS from MRI. To assist the network’s learning in EPVS delineation, we inputted multiple MRI modality sequences to diversify the radiographic features of the EPVS provided for network training. We evaluated our method using the in-house dataset. Results show that the predicted segmentation from our network does not strictly agree with the ground truth on a pixel-by-pixel basis, however, that our method can well identify true EPVSs and has the potential for the clinical purposes of EPVS detection and quantification.