Neural Radiance Field (NeRF) is a promising deep learning approach for three-dimensional (3D) scene reconstruction and view synthesis, with various applications in fields like robotics and medical imaging. However, similar to other deep learning models, understanding NeRF model inaccuracies and their causes is challenging. The 3D nature of NeRFs further adds challenges such as identifying complex geometrical features and analyzing 2D views that suffer from object occlusions. Existing methods for uncertainty quantification (UQ) in NeRFs address the lack of NeRF model understanding by expressing uncertainty in model predictions, exposing limitations in model design or training data. However, these UQ techniques typically rely on quantitative evaluation that does not facilitate human interpretation. We introduce NeRVis, a visual analytics system that supports model users to explore and analyze uncertainty in NeRF scenes. NeRVis combines spatial uncertainty analysis with per-view uncertainty summaries, fostering analysis of the uncertainty in Lambertian NeRF scenes. As a proof-of-concept, we illustrate our approach using two UQ methods. We demonstrate the effectiveness of NeRVis with two different use scenarios, tackling key challenges in the NeRF UQ literature.

Since the various contrast-weighted MR images of a given anatomy contain redundant information, one contrast can be used to guide the reconstruction of another undersampled contrast acquired subsequently in the same session. To solve this reconstruction problem leveraging multi-contrast side information, several end-to-end learning-based guided reconstruction methods have been proposed. However, a key challenge is the requirement for large paired training datasets comprising raw k-space data and aligned reference images. We propose a modular plug-and-play approach, which requires no k-space training data and relies solely on partially paired image-domain datasets. In this approach, a content/style model of two-contrast MR data is first learned from a purely image-domain dataset and subsequently applied as a plug-and-play operator in iterative reconstruction. The disentanglement of content and style allows explicit representation of contrast-independent and contrast-specific factors. Consequently, incorporating prior information into the reconstruction reduces to a simple replacement operation on the aliased content of the estimated image using high-quality content derived from the reference scan. Combining this so-called content consistency operation with an MR data consistency step, followed by a corrective procedure for the content estimate, yields an iterative scheme. We name this novel approach PnP-CoSMo. This approach, by design, offers cross-contrast generalizability and provides an explanatory framework based on the shared and non-shared generative factors underlying the two given contrasts. We explore various aspects of PnP-CoSMo, including interpretability and convergence, via simulations. Furthermore, its practicality is demonstrated on the public NYU fastMRI DICOM dataset, showing equivalent or superior quality and greater generalizability compared to end-to-end methods. On two in-house multi-coil datasets, PnP-CoSMo enabled up to 32.6% greater acceleration over non-guided plug-and-play reconstruction at given SSIM.

Dynamic three-dimensional (4D) reconstruction from two-dimensional X-ray coronary angiography (CA) remains a significant clinical problem. Existing CA reconstruction methods often require extensive user interaction or large training datasets. Recently, Neural Radiance Field (NeRF) has successfully reconstructed high-fidelity scenes in natural and medical contexts without these requirements. However, challenges such as sparse-views, intra-scan motion, and complex vessel morphology hinder its direct application to CA data. We introduce NeRF-CA, a first step toward a fully automatic 4D CA reconstruction that achieves reconstructions from sparse coronary angiograms. To the best of our knowledge, we are the first to address the challenges of sparse-views and cardiac motion by decoupling the scene into the moving coronary artery and the static background, effectively translating the problem of motion into a strength. NeRF-CA serves as a first stepping stone for solving the 4D CA reconstruction problem, achieving adequate 4D reconstructions from as few as four angiograms, as required by clinical practice, while significantly outperforming state-of-the-art sparse-view X-ray NeRF. We validate our approach quantitatively and qualitatively using representative 4D phantom datasets and ablation studies.

Data features and class probabilities are two main perspectives when, e.g., evaluating model results and identifying problematic items. Class probabilities represent the likelihood that each instance belongs to a particular class, which can be produced by probabilistic classifiers or even human labeling with uncertainty. Since both perspectives are multi-dimensional data, dimensionality reduction (DR) techniques are commonly used to extract informative characteristics from them. However, existing methods either focus solely on the data feature perspective or rely on class probability estimates to guide the DR process. In contrast to previous work where separate views are linked to conduct the analysis, we propose a novel approach, class-constrained t-SNE, that combines data features and class probabilities in the same DR result. Specifically, we combine them by balancing two corresponding components in a cost function to optimize the positions of data points and iconic representation of classes - class landmarks. Furthermore, an interactive user-adjustable parameter balances these two components so that users can focus on the weighted perspectives of interest and also empowers a smooth visual transition between varying perspectives to preserve the mental map. We illustrate its application potential in model evaluation and visual-interactive labeling. A comparative analysis is performed to evaluate the DR results.

In recent years, visual analytics (VA) has shown promise in alleviating the challenges of interpreting black-box deep learning (DL) models. While the focus of VA for explainable DL has been mainly on classification problems, DL is gaining popularity in high-dimensional-to-high-dimensional (H-H) problems such as image-to-image translation. In contrast to classification, H-H problems have no explicit instance groups or classes to study. Each output is continuous, high-dimensional, and changes in an unknown non-linear manner with changes in the input. These unknown relations between the input, model and output necessitate the user to analyze them in conjunction, leveraging symmetries between them. Since classification tasks do not exhibit some of these challenges, most existing VA systems and frameworks allow limited control of the components required to analyze models beyond classification. Hence, we identify the need for and present a unified conceptual framework, the Transform-and-Perform framework (T&P), to facilitate the design of VA systems for DL model analysis focusing on H-H problems. T&P provides a checklist to structure and identify workflows and analysis strategies to design new VA systems, and understand existing ones to uncover potential gaps for improvements. The goal is to aid the creation of effective VA systems that support the structuring of model understanding and identifying actionable insights for model improvements. We highlight the growing need for new frameworks like T&P with a real-world image-to-image translation application. We illustrate how T&P effectively supports the understanding and identification of potential gaps in existing VA systems.

Deep learning (DL) models have shown performance benefits across many applications, from classification to image-to-image translation. However, low interpretability often leads to unexpected model behavior once deployed in the real world. Usually, this unexpected behavior is because the training data domain does not reflect the deployment data domain. Identifying a model's breaking points under input conditions and domain shifts, i.e., input transformations, is essential to improve models. Although visual analytics (VA) has shown promise in studying the behavior of model outputs under continually varying inputs, existing methods mainly focus on per-class or instance-level analysis. We aim to generalize beyond classification where classes do not exist and provide a global view of model behavior under co-occurring input transformations. We present a DL model-agnostic VA method (ProactiV) to help model developers proactively study output behavior under input transformations to identify and verify breaking points. ProactiV relies on a proposed input optimization method to determine the changes to a given transformed input to achieve the desired output. The data from this optimization process allows the study of global and local model behavior under input transformations at scale. Additionally, the optimization method provides insights into the input characteristics that result in desired outputs and helps recognize model biases. We highlight how ProactiV effectively supports studying model behavior with example classification and image-to-image translation tasks.

Subsampling a signal of interest can reduce costly data transfer, battery drain, radiation exposure and acquisition time in a wide range of problems. The recently proposed Deep Probabilistic Subsampling (DPS) method effectively integrates subsampling in an end-to-end deep learning model, but learns a static pattern for all datapoints. We generalize DPS to a sequential method that actively picks the next sample based on the information acquired so far; dubbed Active-DPS (A-DPS). We validate that A-DPS improves over DPS for MNIST classification at high subsampling rates. Moreover, we demonstrate strong performance in active acquisition Magnetic Resonance Image (MRI) reconstruction, outperforming DPS and other deep learning methods.

In recent years the t-distributed Stochastic Neighbor Embedding (t-SNE) algorithm has become one of the most used and insightful techniques for exploratory data analysis of high-dimensional data. It reveals clusters of high-dimensional data points at different scales while only requiring minimal tuning of its parameters. However, the computational complexity of the algorithm limits its application to relatively small datasets. To address this problem, several evolutions of t-SNE have been developed in recent years, mainly focusing on the scalability of the similarity computations between data points. However, these contributions are insufficient to achieve interactive rates when visualizing the evolution of the t-SNE embedding for large datasets. In this work, we present a novel approach to the minimization of the t-SNE objective function that heavily relies on graphics hardware and has linear computational complexity. Our technique decreases the computational cost of running t-SNE on datasets by orders of magnitude and retains or improves on the accuracy of past approximated techniques. We propose to approximate the repulsive forces between data points by splatting kernel textures for each data point. This approximation allows us to reformulate the t-SNE minimization problem as a series of tensor operations that can be efficiently executed on the graphics card. An efficient implementation of our technique is integrated and available for use in the widely used Google TensorFlow.js, and an open-source C++ library.

Abstract Hierarchical embeddings, such as HSNE, address critical visual and computational scalability issues of traditional techniques for dimensionality reduction. The improved scalability comes at the cost of the need for increased user interaction for exploration. In this paper, we provide a solution for the interactive visual Focus+Context exploration of such embeddings. We explain how to integrate embedding parts from different levels of detail, corresponding to focus and context groups, in a joint visualization. We devise an according interaction model that relates typical semantic operations on a Focus+Context visualization with the according changes in the level‐of‐detail‐hierarchy of the embedding, including also a mode for comparative Focus+Context exploration and extend HSNE to incorporate the presented interaction model. In order to demonstrate the effectiveness of our approach, we present a use case based on the visual exploration of multi‐dimensional images.

Abstract A bipartite graph is a powerful abstraction for modeling relationships between two collections. Visualizations of bipartite graphs allow users to understand the mutual relationships between the elements in the two collections, e.g., by identifying clusters of similarly connected elements. However, commonly‐used visual representations do not scale for the analysis of large bipartite graphs containing tens of millions of vertices, often resorting to an a‐priori clustering of the sets. To address this issue, we present the Who's‐Active‐On‐What‐Visualization (WAOW‐Vis) that allows for multiscale exploration of a bipartite social‐network without imposing an a‐priori clustering. To this end, we propose to treat a bipartite graph as a high‐dimensional space and we create the WAOW‐Vis adapting the multiscale dimensionality‐reduction technique HSNE. The application of HSNE for bipartite graph requires several modifications that form the contributions of this work. Given the nature of the problem, a set‐based similarity is proposed. For efficient and scalable computations, we use compressed bitmaps to represent sets and we present a novel space partitioning tree to efficiently compute similarities; the Sets Intersection Tree. Finally, we validate WAOW‐Vis on several datasets connecting Twitter‐users and ‐streams in different domains: news, computer science and politics. We show how WAOW‐Vis is particularly effective in identifying hierarchies of communities among social‐media users.

Single-cell analysis through mass cytometry has become an increasingly important tool for immunologists to study the immune system in health and disease. Mass cytometry creates a high-dimensional description vector for single cells by time-of-flight measurement. Recently, t-Distributed Stochastic Neighborhood Embedding (t-SNE) has emerged as one of the state-of-the-art techniques for the visualization and exploration of single-cell data. Ever increasing amounts of data lead to the adoption of Hierarchical Stochastic Neighborhood Embedding (HSNE), enabling the hierarchical representation of the data. Here, the hierarchy is explored selectively by the analyst, who can request more and more detail in areas of interest. Such hierarchies are usually explored by visualizing disconnected plots of selections in different levels of the hierarchy. This poses problems for navigation, by imposing a high cognitive load on the analyst. In this work, we present an interactive summary-visualization to tackle this problem. CyteGuide guides the analyst through the exploration of hierarchically represented single-cell data, and provides a complete overview of the current state of the analysis. We conducted a two-phase user study with domain experts that use HSNE for data exploration. We first studied their problems with their current workflow using HSNE and the requirements to ease this workflow in a field study. These requirements have been the basis for our visual design. In the second phase, we verified our proposed solution in a user evaluation.

Deep neural networks are now rivaling human accuracy in several pattern recognition problems. Compared to traditional classifiers, where features are handcrafted, neural networks learn increasingly complex features directly from the data. Instead of handcrafting the features, it is now the network architecture that is manually engineered. The network architecture parameters such as the number of layers or the number of filters per layer and their interconnections are essential for good performance. Even though basic design guidelines exist, designing a neural network is an iterative trial-and-error process that takes days or even weeks to perform due to the large datasets used for training. In this paper, we present DeepEyes, a Progressive Visual Analytics system that supports the design of neural networks during training. We present novel visualizations, supporting the identification of layers that learned a stable set of patterns and, therefore, are of interest for a detailed analysis. The system facilitates the identification of problems, such as superfluous filters or layers, and information that is not being captured by the network. We demonstrate the effectiveness of our system through multiple use cases, showing how a trained network can be compressed, reshaped and adapted to different problems.

Progressive Visual Analytics aims at improving the interactivity in existing analytics techniques by means of visualization as well as interaction with intermediate results. One key method for data analysis is dimensionality reduction, for example, to produce 2D embeddings that can be visualized and analyzed efficiently. t-Distributed Stochastic Neighbor Embedding (tSNE) is a well-suited technique for the visualization of high-dimensional data. tSNE can create meaningful intermediate results but suffers from a slow initialization that constrains its application in Progressive Visual Analytics. We introduce a controllable tSNE approximation (A-tSNE), which trades off speed and accuracy, to enable interactive data exploration. We offer real-time visualization techniques, including a density-based solution and a Magic Lens to inspect the degree of approximation. With this feedback, the user can decide on local refinements and steer the approximation level during the analysis. We demonstrate our technique with several datasets, in a real-world research scenario and for the real-time analysis of high-dimensional streams to illustrate its effectiveness for interactive data analysis.

Abstract To understand how the immune system works, one needs to have a clear picture of its cellular compositon and the cells' corresponding properties and functionality. Mass cytometry is a novel technique to determine the properties of single‐cells with unprecedented detail. This amount of detail allows for much finer differentiation but also comes at the cost of more complex analysis. In this work, we present Cytosplore, implementing an interactive workflow to analyze mass cytometry data in an integrated system, providing multiple linked views, showing different levels of detail and enabling the rapid definition of known and unknown cell types. Cytosplore handles millions of cells, each represented as a high‐dimensional data point, facilitates hypothesis generation and confirmation, and provides a significant speed up of the current workflow. We show the effectiveness of Cytosplore in a case study evaluation.

Abstract In recent years, dimensionality‐reduction techniques have been developed and are widely used for hypothesis generation in Exploratory Data Analysis. However, these techniques are confronted with overcoming the trade‐off between computation time and the quality of the provided dimensionality reduction. In this work, we address this limitation, by introducing Hierarchical Stochastic Neighbor Embedding (Hierarchical‐SNE). Using a hierarchical representation of the data, we incorporate the well‐known mantra of Overview‐First, Details‐On‐Demand in non‐linear dimensionality reduction. First, the analysis shows an embedding, that reveals only the dominant structures in the data (Overview). Then, by selecting structures that are visible in the overview, the user can filter the data and drill down in the hierarchy. While the user descends into the hierarchy, detailed visualizations of the high‐dimensional structures will lead to new insights. In this paper, we explain how Hierarchical‐SNE scales to the analysis of big datasets. In addition, we show its application potential in the visualization of Deep‐Learning architectures and the analysis of hyperspectral images.