Purpose: Most of the existing convolutional neural network (CNN)-based medical image segmentation methods are based on methods that have originally been developed for segmentation of natural images. Therefore, they largely ignore the differences between the two domains, such as the smaller degree of variability in the shape and appearance of the target volume and the smaller amounts of training data in medical applications. We propose a CNN-based method for prostate segmentation in MRI that employs statistical shape models to address these issues.
Methods: Our CNN predicts the location of the prostate center and the parameters of the shape model, which determine the position of prostate surface keypoints. To train such a large model for segmentation of 3D images using small data (1) we adopt a stage-wise training strategy by first training the network to predict the prostate center and subsequently adding modules for predicting the parameters of the shape model and prostate rotation, (2) we propose a data augmentation method whereby the training images and their prostate surface keypoints are deformed according to the displacements computed based on the shape model, and (3) we employ various regularization techniques.
Results: Our proposed method achieves a Dice score of 0.88, which is obtained by using both elastic-net and spectral dropout for regularization. Compared with a standard CNN-based method, our method shows significantly better segmentation performance on the prostate base and apex. Our experiments also show that data augmentation using the shape model significantly improves the segmentation results.
Conclusions: Prior knowledge about the shape of the target organ can improve the performance of CNN-based segmentation methods, especially where image features are not sufficient for a precise segmentation. Statistical shape models can also be employed to synthesize additional training data that can ease the training of large CNNs.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1007/s11548-018-1785-8 | DOI Listing |
J Imaging Inform Med
January 2025
Faculty of Mathematics and Information Science, Warsaw University of Technology, Warsaw, Poland.
Analysis of the symmetry of the brain hemispheres at the level of individual structures and dominant tissue features has been the subject of research for many years in the context of improving the effectiveness of imaging methods for the diagnosis of brain tumor, stroke, and Alzheimer's disease, among others. One useful approach is to reliably determine the midline of the brain, which allows comparative analysis of the hemispheres and uncovers information on symmetry/asymmetry in the relevant planes of, for example, CT scans. Therefore, an effective method that is robust to various geometric deformations, artifacts, varying noise characteristics, and natural anatomical variability is sought.
View Article and Find Full Text PDFMicrob Ecol
January 2025
State Environmental Protection Key Laboratory of Integrated Surface Water-Groundwater Pollution Control, School of Environmental Science and Engineering, Southern University of Science and Technology, Shenzhen, 518055, Guangdong, China.
The ecological niche separation of microbial interactions in forest ecosystems is critical to maintaining ecological balance and biodiversity and has yet to be comprehensively explored in microbial ecology. This study investigated the impacts of soil properties on microbial interactions and carbon metabolism potential in forest soils across 67 sites in China. Using redundancy analysis and random forest models, we identified soil pH and dissolved organic matter (DOM) aromaticity as the primary drivers of microbial interactions, representing abiotic conditions and resource niches, respectively.
View Article and Find Full Text PDFCommun Biol
January 2025
Department of Mathematical and Statistical Sciences, Faculty of Science, University of Alberta, Edmonton, AB, Canada.
The beauty of tulips has enchanted mankind for centuries. The striped variety has attracted particular attention for its intricate and unpredictable patterns. A good understanding of the mechanism driving the striped pattern formation of broken tulips has been missing since the 17th century.
View Article and Find Full Text PDFSci Rep
January 2025
College of Engineering, Ocean University of China, Qingdao, 266404, China.
Although deterministic analysis can provide initial insights into slope stability, there is no way to reflect the true distribution of soil properties within a slope. To further investigate the effects of the spatial variability of soil properties on the stability and failure mechanism of slope under different foundation types, this study develops a framework combining simple limit equilibrium method (LEM), Monte Carlo Simulation (MCS), and random field to incorporate these factors into slope probabilistic stability analysis. The slope models of two typical foundations (e.
View Article and Find Full Text PDFSci Rep
January 2025
Department of Health Administration, Yonsei University Graduate School, Wonju, Republic of Korea.
This study is the first to examine the determinants of future anxiety in South Korea using the Social Ecological Model (SEM). It aimed to show that, beyond individual factors, mezzo- and macro-level aspects, particularly those related to housing, may influence anxiety. Utilizing 2018 data from the Korean Health Panel Survey, we employed a three-level multilevel analysis to investigate how these factors contribute to the perception of future anxiety among Koreans.
View Article and Find Full Text PDFEnter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!