Search

Background: The most common dermatological complication of insulin therapy is lipohypertrophy. Objective: As a proof of concept, we built and tested an automated model using a convolutional neural network (CNN) to detect the presence of lipohypertrophy in ultrasound images. Methods: Ultrasound images were obtained in a blinded fashion using a portable GE LOGIQ e machine with an L8-18I-D probe (5-18 MHz; GE Healthcare). The data were split into train, validation, and test splits of 70%, 15%, and 15%, respectively. Given the small size of the data set, image augmentation techniques were used to expand the size of the training set and improve the model’s generalizability. To compare the performance of the different architectures, the team considered the accuracy and recall of the models when tested on our test set. Results: The DenseNet CNN architecture was found to have the highest accuracy (76%) and recall (76%) in detecting lipohypertrophy in ultrasound images compared to other CNN architectures. Additional work showed that the YOLOv5m object detection model could be used to help detect the approximate location of lipohypertrophy in ultrasound images identified as containing lipohypertrophy by the DenseNet CNN. Conclusions: We were able to demonstrate the ability of machine learning approaches to automate the process of detecting and locating lipohypertrophy. [ABSTRACT FROM AUTHOR]

Classifying images using supervised machine learning (ML) relies on labeled training data—classes or text descriptions, for example, associated with each image. Data‐driven models are only as good as the data used for training, and this points to the importance of high‐quality labeled data for developing a ML model that has predictive skill. Labeling data is typically a time‐consuming, manual process. Here, we investigate the process of labeling data, with a specific focus on coastal aerial imagery captured in the wake of hurricanes that affected the Atlantic and Gulf Coasts of the United States. The imagery data set is a rich observational record of storm impacts and coastal change, but the imagery requires labeling to render that information accessible. We created an online interface that served labelers a stream of images and a fixed set of questions. A total of 1,600 images were labeled by at least two or as many as seven coastal scientists. We used the resulting data set to investigate interrater agreement: the extent to which labelers labeled each image similarly. Interrater agreement scores, assessed with percent agreement and Krippendorff's alpha, are higher when the questions posed to labelers are relatively simple, when the labelers are provided with a user manual, and when images are smaller. Experiments in interrater agreement point toward the benefit of multiple labelers for understanding the uncertainty in labeling data for machine learning research. Plain Language Summary: After hurricanes and storms, pictures taken from a plane can be used to observe how the coast was impacted. A single flight might take thousands of pictures. If a computer could automatically analyze the pictures, then a person would not need to look at them one‐by‐one. To teach a computer to analyze images, we need many pictures and many labels that describe what is visible in each picture. But where do we get those labels? Typically, a coastal scientist labels the pictures by sorting them into folders or typing codes into a spreadsheet. But does every coastal scientist label pictures the same way? Some labeling questions are easy to answer, and scientists mostly agree ("Is this image all water?"). Other labeling questions are harder to answer, and cause disagreement ("Was there damage to buildings?"). This paper is about how well scientists agree when labeling the same pictures, and how we can improve agreement among scientists. We try some experiments and offer a few ideas on how to improve agreement. We suggest writing very clear questions, using smaller images, and having a comprehensive manual. It turns out that having a manual with examples—and reading the manual!—really helps. Key Points: We measure agreement among coastal scientists labeling the same sets of poststorm imagesCoastal scientists agree more when rating landforms, less when labeling inferred processesIterating on questions, providing documentation, and using smaller image sizes all increase agreement [ABSTRACT FROM AUTHOR]

After decades of study and significant data collection of time-varying swash on sandy beaches, there is no single deterministic prediction scheme for wave runup that eliminates prediction error – even bespoke, locally tuned predictors present scatter when compared to observations. Scatter in runup prediction is meaningful and can be used to create probabilistic predictions of runup for a given wave climate and beach slope. This contribution demonstrates this using a data-driven Gaussian process predictor; a probabilistic machine-learning technique. The runup predictor is developed using 1 year of hourly wave runup data (8328 observations) collected by a fixed lidar at Narrabeen Beach, Sydney, Australia. The Gaussian process predictor accurately predicts hourly wave runup elevation when tested on unseen data with a root-mean-squared error of 0.18 m and bias of 0.02 m. The uncertainty estimates output from the probabilistic GP predictor are then used practically in a deterministic numerical model of coastal dune erosion, which relies on a parameterization of wave runup, to generate ensemble predictions. When applied to a dataset of dune erosion caused by a storm event that impacted Narrabeen Beach in 2011, the ensemble approach reproduced ∼85 % of the observed variability in dune erosion along the 3.5 km beach and provided clear uncertainty estimates around these predictions. This work demonstrates how data-driven methods can be used with traditional deterministic models to develop ensemble predictions that provide more information and greater forecasting skill when compared to a single model using a deterministic parameterization – an idea that could be applied more generally to other numerical models of geomorphic systems. [ABSTRACT FROM AUTHOR]

Abstract Bayesian Networks (BNs) are useful methods of probabilistically modelling environmental systems. BN performance is sensitive to the number of variables included in the model framework. The selection of the optimum set of variables to include in a BN ("variable selection") is therefore a key part of the BN modelling process. While variable selection is an issue dealt with in the wider BN and machine learning literature, it remains largely absent from environmental BN applications to date, due in large part to a lack of software designed to work with available BN packages. CVNetica_VS is an open-source Python module that extends the functionality of Netica, a commonly used commercial BN software package, to perform variable selection. CVNetica_VS uses wrapper-based variable selection and cross-validation to search for the optimum variable set to use in a BN. The software will aid in objectifying and automating the development of BNs in environmental applications. Highlights • Bayesian Networks (BNs) are useful methods of probabilistically modelling environmental systems. • The process of variable selection aids in developing parsimonious and skilful BNs. • CVNetica_VS is a python tool for performing variable selection in BNs. • CVNetica_VS and the variable selection process are demonstrated with a BN that models coastal erosion. [ABSTRACT FROM AUTHOR]

Bayesian Networks (BNs) are an increasingly popular method for modelling environmental systems. The discretization of continuous variables is often required to use BNs. There are three main methods of discretization; manual, unsupervised, and supervised. Here, we compare and demonstrate each approach with a BN that predicts coastal erosion. Results reveal that supervised discretization methods produced BNs of the highest average predictive skill (73.8%), followed by manual discretization (69.0%) and unsupervised discretization (64.8%). However, each method has specific advantages that may make them more suitable for particular applications. Manual methods can produce physical meaningful BNs, which is favorable in environmental modelling. Supervised methods can autonomously and optimally discretize variables and may be preferred when predictive skill is a modelling priority. Unsupervised methods are computationally simple and versatile. The optimal discretization scheme should consider both the performance and practicality of the scheme. [ABSTRACT FROM AUTHOR]

Thermal infrared (IR) imagery is combined with in situ flow measurements to examine the impact of subsurface stratification on boil activity in the tidally influenced Snohomish River. Boils at the river’s surface are an expression of bottom-generated turbulence, appearing as a disruption of the cool-skin surface layer in the IR imagery. Boil activity has previously been linked to the amount of aeration occurring in river systems. A synthesis of data across an ebb tide showed that when a tidal salinity intrusion retreated, turbulent kinetic energy and dissipation rapidly increased by 700% and 575%, respectively. Additionally, the mean boil area fraction in the IR field of view increased by almost 500% across the entire ebb tide time series, with approximately half of this change occurring during the period when the stratification ceased. Using an empirical method for estimating aeration, the change in areal fraction associated with the loss of density stratification is predicted to generate a more than 300% increase in the air-water gas flux. [ABSTRACT FROM PUBLISHER]