Semantic occupancy has recently gained significant traction as a prominent 3D scene representation. However, most existing methods rely on large and costly datasets with fine-grained 3D voxel labels for training, which limits their practicality and scalability, increasing the need for self-monitored learning in this domain. In this work, we present a novel approach to occupancy estimation inspired by neural radiance field (NeRF) using only 2D labels, which are considerably easier to acquire. In particular, we employ differentiable volumetric rendering to predict depth and semantic maps and train a 3D network based on 2D supervision only. To enhance geometric accuracy and increase the supervisory signal, we introduce temporal rendering of adjacent time steps. Additionally, we introduce occupancy flow as a mechanism to handle dynamic objects in the scene and ensure their temporal consistency. Through extensive experimentation we demonstrate that 2D supervision only is sufficient to achieve state-of-the-art performance compared to methods using 3D labels, while outperforming concurrent 2D approaches. When combining 2D supervision with 3D labels, temporal rendering and occupancy flow we outperform all previous occupancy estimation models significantly. We conclude that the proposed rendering supervision and occupancy flow advances occupancy estimation and further bridges the gap towards self-supervised learning in this domain.
Acute Kidney Injury (AKI) is a frequent complication in critically ill patients, affecting up to 50% of patients in the intensive care units. The lack of standardized and open-source tools for applying the Kidney Disease Improving Global Outcomes (KDIGO) criteria to time series data has a negative impact on workload and study quality. This project introduces pyAKI, an open-source pipeline addressing this gap by providing a comprehensive solution for consistent KDIGO criteria implementation. The pyAKI pipeline was developed and validated using a subset of the Medical Information Mart for Intensive Care (MIMIC)-IV database, a commonly used database in critical care research. We defined a standardized data model in order to ensure reproducibility. Validation against expert annotations demonstrated pyAKI's robust performance in implementing KDIGO criteria. Comparative analysis revealed its ability to surpass the quality of human labels. This work introduces pyAKI as an open-source solution for implementing the KDIGO criteria for AKI diagnosis using time series data with high accuracy and performance.
The recently proposed optimization algorithm for deep neural networks Sharpness Aware Minimization (SAM) suggests perturbing parameters before gradient calculation by a gradient ascent step to guide the optimization into parameter space regions of flat loss. While significant generalization improvements and thus reduction of overfitting could be demonstrated, the computational costs are doubled due to the additionally needed gradient calculation, making SAM unfeasible in case of limited computationally capacities. Motivated by Nesterov Accelerated Gradient (NAG) we propose Momentum-SAM (MSAM), which perturbs parameters in the direction of the accumulated momentum vector to achieve low sharpness without significant computational overhead or memory demands over SGD or Adam. We evaluate MSAM in detail and reveal insights on separable mechanisms of NAG, SAM and MSAM regarding training optimization and generalization. Code is available at https://github.com/MarlonBecker/MSAM.
Brain extraction in magnetic resonance imaging (MRI) data is an important segmentation step in many neuroimaging preprocessing pipelines. Image segmentation is one of the research fields in which deep learning had the biggest impact in recent years enabling high precision segmentation with minimal compute. Consequently, traditional brain extraction methods are now being replaced by deep learning-based methods. Here, we used a unique dataset comprising 568 T1-weighted (T1w) MR images from 191 different studies in combination with cutting edge deep learning methods to build a fast, high-precision brain extraction tool called deepbet. deepbet uses LinkNet, a modern UNet architecture, in a two stage prediction process. This increases its segmentation performance, setting a novel state-of-the-art performance during cross-validation with a median Dice score (DSC) of 99.0% on unseen datasets, outperforming current state of the art models (DSC = 97.8% and DSC = 97.9%). While current methods are more sensitive to outliers, resulting in Dice scores as low as 76.5%, deepbet manages to achieve a Dice score of > 96.9% for all samples. Finally, our model accelerates brain extraction by a factor of ~10 compared to current methods, enabling the processing of one image in ~2 seconds on low level hardware.
The brain-age gap is one of the most investigated risk markers for brain changes across disorders. While the field is progressing towards large-scale models, recently incorporating uncertainty estimates, no model to date provides the single-subject risk assessment capability essential for clinical application. In order to enable the clinical use of brain-age as a biomarker, we here combine uncertainty-aware deep Neural Networks with conformal prediction theory. This approach provides statistical guarantees with respect to single-subject uncertainty estimates and allows for the calculation of an individual's probability for accelerated brain-aging. Building on this, we show empirically in a sample of N=16,794 participants that 1. a lower or comparable error as state-of-the-art, large-scale brain-age models, 2. the statistical guarantees regarding single-subject uncertainty estimation indeed hold for every participant, and 3. that the higher individual probabilities of accelerated brain-aging derived from our model are associated with Alzheimer's Disease, Bipolar Disorder and Major Depressive Disorder.
Data acquisition in animal ecology is rapidly accelerating due to inexpensive and accessible sensors such as smartphones, drones, satellites, audio recorders and bio-logging devices. These new technologies and the data they generate hold great potential for large-scale environmental monitoring and understanding, but are limited by current data processing approaches which are inefficient in how they ingest, digest, and distill data into relevant information. We argue that machine learning, and especially deep learning approaches, can meet this analytic challenge to enhance our understanding, monitoring capacity, and conservation of wildlife species. Incorporating machine learning into ecological workflows could improve inputs for population and behavior models and eventually lead to integrated hybrid modeling tools, with ecological models acting as constraints for machine learning models and the latter providing data-supported insights. In essence, by combining new machine learning approaches with ecological domain knowledge, animal ecologists can capitalize on the abundance of data generated by modern sensor technologies in order to reliably estimate population abundances, study animal behavior and mitigate human/wildlife conflicts. To succeed, this approach will require close collaboration and cross-disciplinary education between the computer science and animal ecology communities in order to ensure the quality of machine learning approaches and train a new generation of data scientists in ecology and conservation.
The deviation between chronological age and age predicted from neuroimaging data has been identified as a sensitive risk-marker of cross-disorder brain changes, growing into a cornerstone of biological age-research. However, Machine Learning models underlying the field do not consider uncertainty, thereby confounding results with training data density and variability. Also, existing models are commonly based on homogeneous training sets, often not independently validated, and cannot be shared due to data protection issues. Here, we introduce an uncertainty-aware, shareable, and transparent Monte-Carlo Dropout Composite-Quantile-Regression (MCCQR) Neural Network trained on N=10,691 datasets from the German National Cohort. The MCCQR model provides robust, distribution-free uncertainty quantification in high-dimensional neuroimaging data, achieving lower error rates compared to existing models across ten recruitment centers and in three independent validation samples (N=4,004). In two examples, we demonstrate that it prevents spurious associations and increases power to detect accelerated brain-aging. We make the pre-trained model publicly available.
Overfitting is one of the fundamental challenges when training convolutional neural networks and is usually identified by a diverging training and test loss. The underlying dynamics of how the flow of activations induce overfitting is however poorly understood. In this study we introduce a perplexity-based sparsity definition to derive and visualise layer-wise activation measures. These novel explainable AI strategies reveal a surprising relationship between activation sparsity and overfitting, namely an increase in sparsity in the feature extraction layers shortly before the test loss starts rising. This tendency is preserved across network architectures and reguralisation strategies so that our measures can be used as a reliable indicator for overfitting while decoupling the network's generalisation capabilities from its loss-based definition. Moreover, our differentiable sparsity formulation can be used to explicitly penalise the emergence of sparsity during training so that the impact of reduced sparsity on overfitting can be studied in real-time. Applying this penalty and analysing activation sparsity for well known regularisers and in common network architectures supports the hypothesis that reduced activation sparsity can effectively improve the generalisation and classification performance. In line with other recent work on this topic, our methods reveal novel insights into the contradicting concepts of activation sparsity and network capacity by demonstrating that dense activations can enable discriminative feature learning while efficiently exploiting the capacity of deep models without suffering from overfitting, even when trained excessively.
Age prediction based on Magnetic Resonance Imaging (MRI) data of the brain is a biomarker to quantify the progress of brain diseases and aging. Current approaches rely on preparing the data with multiple preprocessing steps, such as registering voxels to a standardized brain atlas, which yields a significant computational overhead, hampers widespread usage and results in the predicted brain-age to be sensitive to preprocessing parameters. Here we describe a 3D Convolutional Neural Network (CNN) based on the ResNet architecture being trained on raw, non-registered T$_ 1$-weighted MRI data of N=10,691 samples from the German National Cohort and additionally applied and validated in N=2,173 samples from three independent studies using transfer learning. For comparison, state-of-the-art models using preprocessed neuroimaging data are trained and validated on the same samples. The 3D CNN using raw neuroimaging data predicts age with a mean average deviation of 2.84 years, outperforming the state-of-the-art brain-age models using preprocessed data. Since our approach is invariant to preprocessing software and parameter choices, it enables faster, more robust and more accurate brain-age modeling.
Overfitting is one of the most common problems when training deep neural networks on comparatively small datasets. Here, we demonstrate that neural network activation sparsity is a reliable indicator for overfitting which we utilize to propose novel targeted sparsity visualization and regularization strategies. Based on these strategies we are able to understand and counteract overfitting caused by activation sparsity and filter correlation in a targeted layer-by-layer manner. Our results demonstrate that targeted sparsity regularization can efficiently be used to regularize well-known datasets and architectures with a significant increase in image classification performance while outperforming both dropout and batch normalization. Ultimately, our study reveals novel insights into the contradicting concepts of activation sparsity and network capacity by demonstrating that targeted sparsity regularization enables salient and discriminative feature learning while exploiting the full capacity of deep models without suffering from overfitting, even when trained excessively.