Quantifying the unknown impact of segmentation uncertainty on image-based simulations.
Ontology highlight
ABSTRACT:
SUBMITTER: Krygier MC
PROVIDER: S-EPMC8440761 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:This study examines the effect of four different image segmentation methods on the precision of the data. Keywords: other
2004-03-16 | GSE1054 | GEO
Project description:Problem. Image biomarker analysis, also known as radiomics, is a tool for tissue characterization and treatment prognosis that relies on routinely acquired clinical images and delineations. Due to the uncertainty in image acquisition, processing, and segmentation (delineation) protocols, radiomics often lack reproducibility. Radiomics harmonization techniques have been proposed as a solution to reduce these sources of uncertainty and/or their influence on the prognostic model performance. A relevant question is how to estimate the protocol-induced uncertainty of a specific image biomarker, what the effect is on the model performance, and how to optimize the model given the uncertainty. Methods. Two non-small cell lung cancer (NSCLC) cohorts, composed of 421 and 240 patients, respectively, were used for training and testing. Per patient, a Monte Carlo algorithm was used to generate three hundred synthetic contours with a surface dice tolerance measure of less than 1.18 mm with respect to the original GTV. These contours were subsequently used to derive 104 radiomic features, which were ranked on their relative sensitivity to contour perturbation, expressed in the parameter η. The top four (low η) and the bottom four (high η) features were selected for two models based on the Cox proportional hazards model. To investigate the influence of segmentation uncertainty on the prognostic model, we trained and tested the setup in 5000 augmented realizations (using a Monte Carlo sampling method); the log-rank test was used to assess the stratification performance and stability of segmentation uncertainty. Results. Although both low and high η setup showed significant testing set log-rank p-values (p = 0.01) in the original GTV delineations (without segmentation uncertainty introduced), in the model with high uncertainty, to effect ratio, only around 30% of the augmented realizations resulted in model performance with p < 0.05 in the test set. In contrast, the low η setup performed with a log-rank p < 0.05 in 90% of the augmented realizations. Moreover, the high η setup classification was uncertain in its predictions for 50% of the subjects in the testing set (for 80% agreement rate), whereas the low η setup was uncertain only in 10% of the cases. Discussion. Estimating image biomarker model performance based only on the original GTV segmentation, without considering segmentation, uncertainty may be deceiving. The model might result in a significant stratification performance, but can be unstable for delineation variations, which are inherent to manual segmentation. Simulating segmentation uncertainty using the method described allows for more stable image biomarker estimation, selection, and model development. The segmentation uncertainty estimation method described here is universal and can be extended to estimate other protocol uncertainties (such as image acquisition and pre-processing).
| S-EPMC8909427 | biostudies-literature
Project description:Fully convolutional neural networks (FCNs), and in particular U-Nets, have achieved state-of-the-art results in semantic segmentation for numerous medical imaging applications. Moreover, batch normalization and Dice loss have been used successfully to stabilize and accelerate training. However, these networks are poorly calibrated i.e. they tend to produce overconfident predictions for both correct and erroneous classifications, making them unreliable and hard to interpret. In this paper, we study predictive uncertainty estimation in FCNs for medical image segmentation. We make the following contributions: 1) We systematically compare cross-entropy loss with Dice loss in terms of segmentation quality and uncertainty estimation of FCNs; 2) We propose model ensembling for confidence calibration of the FCNs trained with batch normalization and Dice loss; 3) We assess the ability of calibrated FCNs to predict segmentation quality of structures and detect out-of-distribution test examples. We conduct extensive experiments across three medical image segmentation applications of the brain, the heart, and the prostate to evaluate our contributions. The results of this study offer considerable insight into the predictive uncertainty estimation and out-of-distribution detection in medical image segmentation and provide practical recipes for confidence calibration. Moreover, we consistently demonstrate that model ensembling improves confidence calibration.
| S-EPMC7704933 | biostudies-literature
Project description:Prediction of subject age from brain anatomical MRI has the potential to provide a sensitive summary of brain changes, indicative of different neurodegenerative diseases. However, existing studies typically neglect the uncertainty of these predictions. In this work we take into account this uncertainty by applying methods of functional data analysis. We propose a penalised functional quantile regression model of age on brain structure with cognitively normal (CN) subjects in the Alzheimer's Disease Neuroimaging Initiative (ADNI), and use it to predict brain age in Mild Cognitive Impairment (MCI) and Alzheimer's Disease (AD) subjects. Unlike the machine learning approaches available in the literature of brain age prediction, which provide only point predictions, the outcome of our model is a prediction interval for each subject.
| S-EPMC7443707 | biostudies-literature
Project description:The ability to automatically segment an image into distinct regions is a critical aspect in many visual processing applications. Because inaccuracies often exist in automatic segmentation, manual segmentation is necessary in some application domains to correct mistakes, such as required in the reconstruction of neuronal processes from microscopic images. The goal of the automated segmentation tool is traditionally to produce the highest-quality segmentation, where quality is measured by the similarity to actual ground truth, so as to minimize the volume of manual correction necessary. Manual correction is generally orders-of-magnitude more time consuming than automated segmentation, often making handling large images intractable. Therefore, we propose a more relevant goal: minimizing the turn-around time of automated/manual segmentation while attaining a level of similarity with ground truth. It is not always necessary to inspect every aspect of an image to generate a useful segmentation. As such, we propose a strategy to guide manual segmentation to the most uncertain parts of segmentation. Our contributions include 1) a probabilistic measure that evaluates segmentation without ground truth and 2) a methodology that leverages these probabilistic measures to significantly reduce manual correction while maintaining segmentation quality.
| S-EPMC3448604 | biostudies-literature
Project description:With inadequate resources to manage the threats facing biodiversity worldwide, achieving projected management outcomes is critical for efficient resource allocation and species recovery. Despite this, conservation plans to mitigate threats rarely articulate the likelihood of management success. Here we develop a general value of information approach to quantify the impact of uncertainty on 20 threatening processes affecting 976 listed species and communities. To our knowledge, this is the most comprehensive quantification of the impacts of uncertainty on threat management. We discover that, on average, removing uncertainty about management effectiveness could triple the gain in persistence achieved by managing under current uncertainty. Management of fire, invasive animals and a plant pathogen are most impeded by uncertainty; management of invasive plants is least impacted. Our results emphasise the tremendous importance of reducing uncertainty about species responses to management, and show that failure to consider management effectiveness wastes resources and impedes species recovery.
| S-EPMC6687751 | biostudies-literature
Project description:Despite the state-of-the-art performance for medical image segmentation, deep convolutional neural networks (CNNs) have rarely provided uncertainty estimations regarding their segmentation outputs, e.g., model (epistemic) and image-based (aleatoric) uncertainties. In this work, we analyze these different types of uncertainties for CNN-based 2D and 3D medical image segmentation tasks at both pixel level and structure level. We additionally propose a test-time augmentation-based aleatoric uncertainty to analyze the effect of different transformations of the input image on the segmentation output. Test-time augmentation has been previously used to improve segmentation accuracy, yet not been formulated in a consistent mathematical framework. Hence, we also propose a theoretical formulation of test-time augmentation, where a distribution of the prediction is estimated by Monte Carlo simulation with prior distributions of parameters in an image acquisition model that involves image transformations and noise. We compare and combine our proposed aleatoric uncertainty with model uncertainty. Experiments with segmentation of fetal brains and brain tumors from 2D and 3D Magnetic Resonance Images (MRI) showed that 1) the test-time augmentation-based aleatoric uncertainty provides a better uncertainty estimation than calculating the test-time dropout-based model uncertainty alone and helps to reduce overconfident incorrect predictions, and 2) our test-time augmentation outperforms a single-prediction baseline and dropout-based multiple predictions.
| S-EPMC6783308 | biostudies-literature
Project description:Recently, clinical radiological research and practice are becoming increasingly quantitative. Further, images continue to increase in size and volume. For quantitative radiology to become practical, it is crucial that image segmentation algorithms and their implementations are rapid and yield practical run time on very large data sets. The purpose of this paper is to present a parallel version of an algorithm that belongs to the family of fuzzy connectedness (FC) algorithms, to achieve an interactive speed for segmenting large medical image data sets.The most common FC segmentations, optimizing an [script-l](∞)-based energy, are known as relative fuzzy connectedness (RFC) and iterative relative fuzzy connectedness (IRFC). Both RFC and IRFC objects (of which IRFC contains RFC) can be found via linear time algorithms, linear with respect to the image size. The new algorithm, P-ORFC (for parallel optimal RFC), which is implemented by using NVIDIA's Compute Unified Device Architecture (CUDA) platform, considerably improves the computational speed of the above mentioned CPU based IRFC algorithm.Experiments based on four data sets of small, medium, large, and super data size, achieved speedup factors of 32.8×, 22.9×, 20.9×, and 17.5×, correspondingly, on the NVIDIA Tesla C1060 platform. Although the output of P-ORFC need not precisely match that of IRFC output, it is very close to it and, as the authors prove, always lies between the RFC and IRFC objects.A parallel version of a top-of-the-line algorithm in the family of FC has been developed on the NVIDIA GPUs. An interactive speed of segmentation has been achieved, even for the largest medical image data set. Such GPU implementations may play a crucial role in automatic anatomy recognition in clinical radiology.
| S-EPMC3537719 | biostudies-other
Project description:BackgroundAutomatic myocardial scar segmentation from late gadolinium enhancement (LGE) images using neural networks promises an alternative to time-consuming and observer-dependent semi-automatic approaches. However, alterations in data acquisition, reconstruction as well as post-processing may compromise network performance. The objective of the present work was to systematically assess network performance degradation due to a mismatch of point-spread function between training and testing data.MethodsThirty-six high-resolution (0.7×0.7×2.0 mm3) LGE k-space datasets were acquired post-mortem in porcine models of myocardial infarction. The in-plane point-spread function and hence in-plane resolution Δx was retrospectively degraded using k-space lowpass filtering, while field-of-view and matrix size were kept constant. Manual segmentation of the left ventricle (LV) and healthy remote myocardium was performed to quantify location and area (% of myocardium) of scar by thresholding (≥ SD5 above remote). Three standard U-Nets were trained on training resolutions Δxtrain = 0.7, 1.2 and 1.7 mm to predict endo- and epicardial borders of LV myocardium and scar. The scar prediction of the three networks for varying test resolutions (Δxtest = 0.7 to 1.7 mm) was compared against the reference SD5 thresholding at 0.7 mm. Finally, a fourth network trained on a combination of resolutions (Δxtrain = 0.7 to 1.7 mm) was tested.ResultsThe prediction of relative scar areas showed the highest precision when the resolution of the test data was identical to or close to the resolution used during training. The median fractional scar errors and precisions (IQR) from networks trained and tested on the same resolution were 0.0 percentage points (p.p.) (1.24 - 1.45), and - 0.5 - 0.0 p.p. (2.00 - 3.25) for networks trained and tested on the most differing resolutions, respectively. Deploying the network trained on multiple resolutions resulted in reduced resolution dependency with median scar errors and IQRs of 0.0 p.p. (1.24 - 1.69) for all investigated test resolutions.ConclusionA mismatch of the imaging point-spread function between training and test data can lead to degradation of scar segmentation when using current U-Net architectures as demonstrated on LGE porcine myocardial infarction data. Training networks on multi-resolution data can alleviate the resolution dependency.
| S-EPMC10981112 | biostudies-literature
Project description:Climate change studies involve complex processes translating coarse climate change projections in locally meaningful terms. We analysed the behaviour of weather generators while downscaling precipitation and air temperature data. With multiple climate indices and alternative weather generators, we directly quantified the uncertainty associated with using weather generators when site specific downscaling is performed. We extracted the influence of weather generators on climate variability at local scale and the uncertainty that could affect impact assessment. For that, we first designed the downscaling experiments with three weather generators (CLIMAK, LARS-WG, WeaGETS) to interpret future projections. Then we assessed the impacts of estimated changes of precipitation and air temperature for a sample of 15 sites worldwide using a rice yield model and an extended set of climate metrics. We demonstrated that the choice of a weather generator in the downscaling process may have a higher impact on crop yield estimates than the climate scenario adopted. Should they be confirmed, these results would indicate that widely accepted outcomes of climate change studies using this downscaling technique need reconsideration.
| S-EPMC6592885 | biostudies-other