In this work, we introduce an anisotropic minimal path model based on a new Riemannian tensor integrating the crossing-adaptive anisotropic radius-lifted tensor field and the front freezing indicator by appearance and path features. The non-local path feature only can be obtained during the geodesic distance computation process by the fast marching method. The predefined criterion derived from path feature is able to steer the front evolution by freezing the point causing high bending of the geodesic to solve the shortcut problem. We performed qualitative and quantitative experiments on synthetic and real images (including retinal vessels, rivers and roads) and compare with the minimal path models with classical anisotropic Riemannian metric and dynamic isotropic metric, which demonstrated the proposed method can detect desired targets from complex tubular tree structures. © 2020

Deep learning is a hot research topic in the field of machine learning methods and applications. Real-value neural networks (Real NNs), especially deep real networks (DRNs), have been widely used in many research fields. In recent years, the deep complex networks (DCNs) and the deep quaternion networks (DQNs) have attracted more and more attentions. The octonion algebra, which is an extension of complex algebra and quaternion algebra, can provide more efficient and compact expressions. This paper constructs a general framework of deep octonion networks (DONs) and provides the main building blocks of DONs such as octonion convolution, octonion batch normalization and octonion weight initialization; DONs are then used in image classification tasks for CIFAR-10 and CIFAR-100 data sets. Compared with the DRNs, the DCNs, and the DQNs, the proposed DONs have better convergence and higher classification accuracy. The success of DONs is also explained by multi-task learning. (C) 2020 Elsevier B.V. All rights reserved.

Accurate segmentation of uterus, uterine fibroids, and spine from MR images is crucial for high intensity focusedultrasound(HIFU) therapybut remains still difficult to achieve because of 1) the large shape and size variations among individuals, 2) the low contrast between adjacent organs and tissues, and 3) the unknown number of uterine fibroids. To tackle this problem, in this paper, we propose a large kernel Encoder-Decoder Network based on a 2Dsegmentationmodel. The use of this large kernel can capturemulti-scale contextsby enlarging the valid receptive field. In addition, a deep multiple atrous convolution block is also employed to enlarge the receptive field and extract denser feature maps. Our approach is compared to both conventional and other deep learning methods and the experimental results conducted on a large dataset show its effectiveness. © 2020 International Union of Crystallography. All rights reserved.

Objectives: This study was to assess the feasibility of a modified multiphasic CT scan protocol combined with homemade software measurements of glomerular filtration rate (CT-GFR) and explore the effect of renal tumor volume on the calculation of CT-GFR.Materials and Methods: Prospective observational study comparing three methods of GFR measurement from February 2017 to December 2017, 91 patients with unilateral renal tumor underwent both a modified multiphasic CT scans of kidney and serum creatinine (Scr) tests preoperatively, of which 15 cases underwent additional radionuclide examination. Total and split CT-GFR, with or without renal tumor, were quantified by the homemade software in early and late renal parenchymal phases, respectively. The volume of renal tumor was quantified by the homemade software. Correlation and difference between CT-GFR and traditional methods of GFR measurement, including estimated GFR (eGFR) from Scr concentration and split GFR using of radionuclide examination (R-GFR), were performed.Results: There is a strong correlation between CT-GFR with renal tumor and eGFR (r = 0.90, p < 0.001) in early renal parenchymal phase. The relative CT-GFR in early renal parenchymal phase was highly correlated with the relative R-GFR (r = 0.88, p < 0.001). Renal tumor volume significantly correlated with the value of CT-GFR that determined by subtracting the CT-GFR measurement without renal tumor from CT-GFR measurement with renal tumor (r = 0.89, p < 0.001).Conclusion: A modified multiphasic CT scan protocol combined with homemade software might be an alternative technique for the evaluation of renal function for the patients with unilateral renal tumor.