Background: Accurate segmentation of tumor targets is critical for maximizing tumor control and minimizing normal tissue toxicity. We proposed a sequential and iterative U-Net (SI-Net) deep learning method to auto-segment the high-risk primary tumor clinical target volume (CTVp1) for treatment planning of nasopharyngeal carcinoma (NPC) radiotherapy. Methods: The SI-Net is a variant of the U-Net architecture. The input of SI-Net includes one CT image, the CTVp1 contour on this image, and the next CT image. The output is the predicted CTVp1 contour on the next CT image. We designed the SI-Net, using the left side to learn the volumetric features and the right to localize the contour on the next image. Two prediction directions, one from inferior to superior (forward direction) and the other from superior to inferior (backward direction), were tested. The performance was compared between the SI-Net and the U-Net using Dice similarity coefficient (DSC), Jaccard index (JI), average surface distance (ASD), and Hausdorff distance (HD) metrics. Results: The DSC and JI values from the forward direction SI-Net model were 5 and 6% higher than those from the U-Net model (0.84 ± 0.04 vs. 0.80 ± 0.05 and 0.74 ± 0.05 vs. 0.69 ± 0.05, p < 0.001). The smaller ASD and HD values also indicated a better performance (2.8 ± 1.0 vs. 3.3 ± 1.0 mm and 8.7 ± 2.5 vs. 9.7 ± 2.7 mm, p < 0.01) for the SI-Net model. For the backward direction SI-Net model, the DSC and JI values were still better than those from the U-Net model ( p < 0.01), although there were no significant differences in ASD and HD. Conclusions: The SI-Net model preserved the continuity between adjacent images and thus improved the segmentation accuracy compared with the conventional U-Net model. This model has potential of improving the efficiency and consistence of CTVp1 contouring for NPC patients.
It is well known that hurricane intensification is often accompanied by continuous contraction of the radius of maximum wind (RMW) and eyewall size. However, a few recent studies have shown rapid and then slow contraction of the RMW/eyewall size prior to the onset and during the early stages of rapid intensification (RI) of hurricanes, respectively, but a steady state in the RMW (S-RMW) and eyewall size during the later stages of RI. In this study, a statistical analysis of S-RMWs associated with rapidly intensifying hurricanes is performed using the extended best-track dataset during 1990-2014 in order to examine how frequently, and at what intensity and size, the S-RMW structure tends to occur. Results show that about 53% of the 139 RI events of 24-h duration associated with 55 rapidly intensifying hurricanes exhibit S-RMWs, and that the percentage of the S-RMW events increases to 69% when RI events are evaluated at 12-h intervals, based on a new RI rate definition of 10 m s 21 (12 h) 21 ; both results satisfy the Student's t tests with confidence levels of over 95%. In general, S-RMWs tend to appear more frequently in more intense storms and when their RMWs are contracted to less than 50 km. This work suggests a new fruitful research area in studying the RI of hurricanes with S-RMWs.
Hurricane Patricia (2015) broke records in both peak intensity and rapid intensification (RI) rate over the eastern Pacific basin. All of the then-operational models predicted less than half of its extraordinary intensity and RI rate, leaving a challenge for numerical modeling studies. In this study, a successful 42-h simulation of Patricia is obtained using a quintuply nested-grid version of the Weather Research and Forecast (WRF) Model with the finest grid size of 333 m. Results show that the WRF Model, initialized with the Global Forecast System Final Analysis data only, could reproduce the track, peak intensity, and many inner-core features, as verified against various observations. In particular, its simulated maximum surface wind of 92 m s−1 is close to the observed 95 m s−1, capturing the unprecedented RI rate of 54 m s−1 (24 h)−1. In addition, the model reproduces an intense warm-cored eye, a small-sized eyewall with a radius of maximum wind of less than 10 km, and the distribution of narrow spiral rainbands. A series of sensitivity simulations is performed to help understand which model configurations are essential to reproducing the extraordinary intensity of the storm. Results reveal that Patricia’s extraordinary development and its many inner-core structures could be reasonably well simulated if ultrahigh horizontal resolution, appropriate model physics, and realistic initial vortex intensity are incorporated. It is concluded that the large-scale conditions (e.g., warm sea surface temperature, weak vertical wind shear, and the moist intertropical convergence zone) and convective organization play important roles in determining the predictability of Patricia’s extraordinary RI and peak intensity.
The main energy source for tropical cyclones (TCs) is from the warm ocean through their boundary layer (BL) processes. Understanding the TC boundary layer (TCBL) structure has become increasingly important in the effort toward developing high-resolution numerical models and improving TC intensity forecasts (
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.