Sune Darkner

Abraham George Smith, Denis Kutnár, Ivan Richter Vogelius et al.

Increased organ at risk segmentation accuracy is required to reduce cost and complications for patients receiving radiotherapy treatment. Some deep learning methods for the segmentation of organs at risk use a two stage process where a localisation network first crops an image to the relevant region and then a locally specialised network segments the cropped organ of interest. We investigate the accuracy improvements brought about by such a localisation stage by comparing to a single-stage baseline network trained on full resolution images. We find that localisation approaches can improve both training time and stability and a two stage process involving both a localisation and organ segmentation network provides a significant increase in segmentation accuracy for the spleen, pancreas and heart from the Medical Segmentation Decathlon dataset. We also observe increased benefits of localisation for smaller organs. Source code that recreates the main results is available at \href{https://github.com/Abe404/localise_to_segment}{this https URL}.

Stefan Sommer, Mads Nielsen, Sune Darkner et al.

To achieve sparse parametrizations that allows intuitive analysis, we aim to represent deformation with a basis containing interpretable elements, and we wish to use elements that have the description capacity to represent the deformation compactly. To accomplish this, we introduce in this paper higher-order momentum distributions in the LDDMM registration framework. While the zeroth order moments previously used in LDDMM only describe local displacement, the first-order momenta that are proposed here represent a basis that allows local description of affine transformations and subsequent compact description of non-translational movement in a globally non-rigid deformation. The resulting representation contains directly interpretable information from both mathematical and modeling perspectives. We develop the mathematical construction of the registration framework with higher-order momenta, we show the implications for sparse image registration and deformation description, and we provide examples of how the parametrization enables registration with a very low number of parameters. The capacity and interpretability of the parametrization using higher-order momenta lead to natural modeling of articulated movement, and the method promises to be useful for quantifying ventricle expansion and progressing atrophy during Alzheimer's disease.

Peidi Xu, Faezeh Moshfeghifar, Torkan Gholamalizadeh et al.

Accurate geometry representation is essential in developing finite element models. Although generally good, deep-learning segmentation approaches with only few data have difficulties in accurately segmenting fine features, e.g., gaps and thin structures. Subsequently, segmented geometries need labor-intensive manual modifications to reach a quality where they can be used for simulation purposes. We propose a strategy that uses transfer learning to reuse datasets with poor segmentation combined with an interactive learning step where fine-tuning of the data results in anatomically accurate segmentations suitable for simulations. We use a modified MultiPlanar UNet that is pre-trained using inferior hip joint segmentation combined with a dedicated loss function to learn the gap regions and post-processing to correct tiny inaccuracies on symmetric classes due to rotational invariance. We demonstrate this robust yet conceptually simple approach applied with clinically validated results on publicly available computed tomography scans of hip joints. Code and resulting 3D models are available at: https://github.com/MICCAI2022-155/AuToSeg}

Jiahao Lu, Chong Yin, Oswin Krause et al.

Feature-based self-explanatory methods explain their classification in terms of human-understandable features. In the medical imaging community, this semantic matching of clinical knowledge adds significantly to the trustworthiness of the AI. However, the cost of additional annotation of features remains a pressing issue. We address this problem by proposing cRedAnno, a data-/annotation-efficient self-explanatory approach for lung nodule diagnosis. cRedAnno considerably reduces the annotation need by introducing self-supervised contrastive learning to alleviate the burden of learning most parameters from annotation, replacing end-to-end training with two-stage training. When training with hundreds of nodule samples and only 1% of their annotations, cRedAnno achieves competitive accuracy in predicting malignancy, meanwhile significantly surpassing most previous works in predicting nodule attributes. Visualisation of the learned space further indicates that the correlation between the clustering of malignancy and nodule attributes coincides with clinical knowledge. Our complete code is open-source available: https://github.com/diku-dk/credanno.

Jiahao Lu, Chong Yin, Kenny Erleben et al.

Recently, attempts have been made to reduce annotation requirements in feature-based self-explanatory models for lung nodule diagnosis. As a representative, cRedAnno achieves competitive performance with considerably reduced annotation needs by introducing self-supervised contrastive learning to do unsupervised feature extraction. However, it exhibits unstable performance under scarce annotation conditions. To improve the accuracy and robustness of cRedAnno, we propose an annotation exploitation mechanism by conducting semi-supervised active learning with sparse seeding and training quenching in the learned semantically meaningful reasoning space to jointly utilise the extracted features, annotations, and unlabelled data. The proposed approach achieves comparable or even higher malignancy prediction accuracy with 10x fewer annotations, meanwhile showing better robustness and nodule attribute prediction accuracy under the condition of 1% annotations. Our complete code is open-source available: https://github.com/diku-dk/credanno.

Xiang Dai, Ilias Chalkidis, Sune Darkner et al.

The recent literature in text classification is biased towards short text sequences (e.g., sentences or paragraphs). In real-world applications, multi-page multi-paragraph documents are common and they cannot be efficiently encoded by vanilla Transformer-based models. We compare different Transformer-based Long Document Classification (TrLDC) approaches that aim to mitigate the computational overhead of vanilla transformers to encode much longer text, namely sparse attention and hierarchical encoding methods. We examine several aspects of sparse attention (e.g., size of local attention window, use of global attention) and hierarchical (e.g., document splitting strategy) transformers on four document classification datasets covering different domains. We observe a clear benefit from being able to process longer text, and, based on our results, we derive practical advice of applying Transformer-based models on long document classification tasks.

Peidi Xu, Niels-Henrik Holstein-Rathlou, Stinne Byrholdt Søgaard et al.

The renal vasculature, acting as a resource distribution network, plays an important role in both the physiology and pathophysiology of the kidney. However, no imaging techniques allow an assessment of the structure and function of the renal vasculature due to limited spatial and temporal resolution. To develop realistic computer simulations of renal function, and to develop new image-based diagnostic methods based on artificial intelligence, it is necessary to have a realistic full-scale model of the renal vasculature. We propose a hybrid framework to build subject-specific models of the renal vascular network by using semi-automated segmentation of large arteries and estimation of cortex area from a micro-CT scan as a starting point, and by adopting the Global Constructive Optimization algorithm for generating smaller vessels. Our results show a statistical correspondence between the reconstructed data and existing anatomical data obtained from a rat kidney with respect to morphometric and hemodynamic parameters.

Faezeh Moshfeghifar, Max Kragballe Nielsen, José D. Tascón-Vidarte et al.

We present a method to generate subject-specific cartilage for the hip joint. Given bone geometry, our approach is agnostic to image modality, creates conforming interfaces, and is well suited for finite element analysis. We demonstrate our method on ten hip joints showing anatomical shape consistency and well-behaved stress patterns. Our method is fast and may assist in large-scale biomechanical population studies of the hip joint when manual segmentation or training data is not feasible.

Rasmus Netterstrøm, Nikolay Kutuzov, Sune Darkner et al.

Tracking single molecules is instrumental for quantifying the transport of molecules and nanoparticles in biological samples, e.g., in brain drug delivery studies. Existing intensity-based localisation methods are not developed for imaging with a scanning microscope, typically used for in vivo imaging. Low signal-to-noise ratios, movement of molecules out-of-focus, and high motion blur on images recorded with scanning two-photon microscopy (2PM) in vivo pose a challenge to the accurate localisation of molecules. Using data-driven models is challenging due to low data volumes, typical for in vivo experiments. We developed a 2PM image simulator to supplement scarce training data. The simulator mimics realistic motion blur, background fluorescence, and shot noise observed in vivo imaging. Training a data-driven model with simulated data improves localisation quality in simulated images and shows why intensity-based methods fail.

Jiahao Lu, Chong Yin, Silvia Ingala et al.

Lung cancer, a leading cause of cancer-related deaths globally, emphasises the importance of early detection for better patient outcomes. Pulmonary nodules, often early indicators of lung cancer, necessitate accurate, timely diagnosis. Despite Explainable Artificial Intelligence (XAI) advances, many existing systems struggle providing clear, comprehensive explanations, especially with limited labelled data. This study introduces MERA, a Multimodal and Multiscale self-Explanatory model designed for lung nodule diagnosis with considerably Reduced Annotation requirements. MERA integrates unsupervised and weakly supervised learning strategies (self-supervised learning techniques and Vision Transformer architecture for unsupervised feature extraction) and a hierarchical prediction mechanism leveraging sparse annotations via semi-supervised active learning in the learned latent space. MERA explains its decisions on multiple levels: model-level global explanations via semantic latent space clustering, instance-level case-based explanations showing similar instances, local visual explanations via attention maps, and concept explanations using critical nodule attributes. Evaluations on the public LIDC dataset show MERA's superior diagnostic accuracy and self-explainability. With only 1% annotated samples, MERA achieves diagnostic accuracy comparable to or exceeding state-of-the-art methods requiring full annotation. The model's inherent design delivers comprehensive, robust, multilevel explanations aligned closely with clinical practice, enhancing trustworthiness and transparency. Demonstrated viability of unsupervised and weakly supervised learning lowers the barrier to deploying diagnostic AI in broader medical domains. Our complete code is open-source available: https://github.com/diku-dk/credanno.

Peidi Xu, Olga Sosnovtseva, Charlotte Mehlin Sørensen et al.

Accurate analysis and modeling of renal functions require a precise segmentation of the renal blood vessels. Micro-CT scans provide image data at higher resolutions, making more small vessels near the renal cortex visible. Although deep-learning-based methods have shown state-of-the-art performance in automatic blood vessel segmentations, they require a large amount of labeled training data. However, voxel-wise labeling in micro-CT scans is extremely time-consuming given the huge volume sizes. To mitigate the problem, we simulate synthetic renal vascular trees physiologically while generating corresponding scans of the simulated trees by training a generative model on unlabeled scans. This enables the generative model to learn the mapping implicitly without the need for explicit functions to emulate the image acquisition process. We further propose an additional segmentation branch over the generative model trained on the generated scans. We demonstrate that the model can directly segment blood vessels on real scans and validate our method on both 3D micro-CT scans of rat kidneys and a proof-of-concept experiment on 2D retinal images. Code and 3D results are available at https://github.com/miccai2023anony/RenalVesselSeg

Abraham George Smith, Jens Petersen, Cynthia Terrones-Campos et al.

Organ-at-risk contouring is still a bottleneck in radiotherapy, with many deep learning methods falling short of promised results when evaluated on clinical data. We investigate the accuracy and time-savings resulting from the use of an interactive-machine-learning method for an organ-at-risk contouring task. We compare the method to the Eclipse contouring software and find strong agreement with manual delineations, with a dice score of 0.95. The annotations created using corrective-annotation also take less time to create as more images are annotated, resulting in substantial time savings compared to manual methods, with hearts that take 2 minutes and 2 seconds to delineate on average, after 923 images have been delineated, compared to 7 minutes and 1 seconds when delineating manually. Our experiment demonstrates that interactive-machine-learning with corrective-annotation provides a fast and accessible way for non computer-scientists to train deep-learning models to segment their own structures of interest as part of routine clinical workflows. Source code is available at \href{https://github.com/Abe404/RootPainter3D}{this HTTPS URL}.

Yibo Liu, Zhixin Fang, Sune Darkner et al.

We propose mesh-free fluid simulations that exploit a kinematic neural basis for velocity fields represented by an MLP. We design a set of losses that ensures that these neural bases approximate fundamental physical properties such as orthogonality, divergence-free, boundary alignment, and smoothness. Our neural bases can then be used to fit an input sketch of a flow, which will inherit the same fundamental properties from the bases. We then can animate such flow in real-time using standard time integrators. Our neural bases can accommodate different domains, moving boundaries, and naturally extend to three dimensions.

Jon Sporring, Peidi Xu, Jiahao Lu et al.

We present Locally Orderless Networks (LON) and its theoretic foundation which links it to Convolutional Neural Networks (CNN), to Scale-space histograms, and measurement theory. The key elements are a regular sampling of the bias and the derivative of the activation function. We compare LON, CNN, and Scale-space histograms on prototypical single-layer networks. We show how LON and CNN can emulate each other, how LON expands the set of functionals computable to non-linear functions such as squaring. We demonstrate simple networks which illustrate the improved performance of LON over CNN on simple tasks for estimating the gradient magnitude squared, for regressing shape area and perimeter lengths, and for explainability of individual pixels' influence on the result.

Sune Darkner, Jose D Tascon, Francois Lauze

First Order Locally Orderless Registration (FLOR) is a scale-space framework for image density estimation used for defining image similarity, mainly for Image Registration. The Locally Orderless Registration framework was designed in principle to use zeroth-order information, providing image density estimates over three scales: image scale, intensity scale, and integration scale. We extend it to take first-order information into account and hint at higher-order information. We show how standard similarity measures extend into the framework. We study especially Sum of Squared Differences (SSD) and Normalized Cross-Correlation (NCC) but present the theory of how Normalised Mutual Information (NMI) can be included.

Mathias Perslev, Michael Hejselbak Jensen, Sune Darkner et al.

Neural networks are becoming more and more popular for the analysis of physiological time-series. The most successful deep learning systems in this domain combine convolutional and recurrent layers to extract useful features to model temporal relations. Unfortunately, these recurrent models are difficult to tune and optimize. In our experience, they often require task-specific modifications, which makes them challenging to use for non-experts. We propose U-Time, a fully feed-forward deep learning approach to physiological time series segmentation developed for the analysis of sleep data. U-Time is a temporal fully convolutional network based on the U-Net architecture that was originally proposed for image segmentation. U-Time maps sequential inputs of arbitrary length to sequences of class labels on a freely chosen temporal scale. This is done by implicitly classifying every individual time-point of the input signal and aggregating these classifications over fixed intervals to form the final predictions. We evaluated U-Time for sleep stage classification on a large collection of sleep electroencephalography (EEG) datasets. In all cases, we found that U-Time reaches or outperforms current state-of-the-art deep learning models while being much more robust in the training process and without requiring architecture or hyperparameter adaptation across tasks.

Henrik Grønholt Jensen, François Lauze, Sune Darkner

We present an information-theoretic approach to the registration of images with directional information, and especially for diffusion-Weighted Images (DWI), with explicit optimization over the directional scale. We call it Locally Orderless Registration with Directions (LORD). We focus on normalized mutual information as a robust information-theoretic similarity measure for DWI. The framework is an extension of the LOR-DWI density-based hierarchical scale-space model that varies and optimizes the integration, spatial, directional, and intensity scales. As affine transformations are insufficient for inter-subject registration, we extend the model to non-rigid deformations. We illustrate that the proposed model deforms orientation distribution functions (ODFs) correctly and is capable of handling the classic complex challenges in DWI-registrations, such as the registration of fiber-crossings along with kissing, fanning, and interleaving fibers. Our experimental results clearly illustrate a novel promising regularizing effect, that comes from the nonlinear orientation-based cost function. We show the properties of the different image scales and, we show that including orientational information in our model makes the model better at retrieving deformations in contrast to standard scalar-based registration.

Sune Darkner, Stefan Sommer, Andreas Schuhmacher et al.

Humans are highly dependent on the ability to process audio in order to interact through conversation and navigate from sound. For this, the shape of the ear acts as a mechanical audio filter. The anatomy of the outer human ear canal to approximately 15-20 mm beyond the Tragus is well described because of its importance for customized hearing aid production. This is however not the case for the part of the ear canal that is embedded in the skull, until the typanic membrane. Due to the sensitivity of the outer ear, this part, referred to as the bony part, has only been described in a few population studies and only ex-vivo. We present a study of the entire ear canal including the bony part and the tympanic membrane. We form an average ear canal from a number of MRI scans using standard image registration methods. We show that the obtained representation is realistic in the sense that it has acoustical properties almost identical to a real ear.

Akshay Pai, Stefan Sommer, Lars Lau Raket et al.

Template estimation plays a crucial role in computational anatomy since it provides reference frames for performing statistical analysis of the underlying anatomical population variability. While building models for template estimation, variability in sites and image acquisition protocols need to be accounted for. To account for such variability, we propose a generative template estimation model that makes simultaneous inference of both bias fields in individual images, deformations for image registration, and variance hyperparameters. In contrast, existing maximum a posterori based methods need to rely on either bias-invariant similarity measures or robust image normalization. Results on synthetic and real brain MRI images demonstrate the capability of the model to capture heterogeneity in intensities and provide a reliable template estimation from registration.

Sune Darkner, Jon Sporring

Image registration is an important tool for medical image analysis and is used to bring images into the same reference frame by warping the coordinate field of one image, such that some similarity measure is minimized. We study similarity in image registration in the context of Locally Orderless Images (LOI), which is the natural way to study density estimates and reveals the 3 fundamental scales: the measurement scale, the intensity scale, and the integration scale. This paper has three main contributions: Firstly, we rephrase a large set of popular similarity measures into a common framework, which we refer to as Locally Orderless Registration, and which makes full use of the features of local histograms. Secondly, we extend the theoretical understanding of the local histograms. Thirdly, we use our framework to compare two state-of-the-art intensity density estimators for image registration: The Parzen Window (PW) and the Generalized Partial Volume (GPV), and we demonstrate their differences on a popular similarity measure, Normalized Mutual Information (NMI). We conclude, that complicated similarity measures such as NMI may be evaluated almost as fast as simple measures such as Sum of Squared Distances (SSD) regardless of the choice of PW and GPV. Also, GPV is an asymmetric measure, and PW is our preferred choice.