Detection of unwanted (‘foreign’) objects within products is a common procedure in many branches of industry for maintaining production quality. X-ray imaging is a fast, non-invasive and widely applicable method for foreign object detection. Deep learning has recently emerged as a powerful approach for recognizing patterns in radiographs (i.e., X-ray images), enabling automated X-ray based foreign object detection. However, these methods require a large number of training examples and manual annotation of these examples is a subjective and laborious task. In this work, we propose a Computed Tomography (CT) based method for producing training data for supervised learning of foreign object detection, with minimal labor requirements. In our approach, a few representative objects are CT scanned and reconstructed in 3D. The radiographs that are acquired as part of the CT-scan data serve as input for the machine learning method. High-quality ground truth locations of the foreign objects are obtained through accurate 3D reconstructions and segmentations. Using these segmented volumes, corresponding 2D segmentations are obtained by creating virtual projections. We outline the benefits of objectively and reproducibly generating training data in this way. In addition, we show how the accuracy depends on the number of objects used for the CT reconstructions. The results show that in this workflow generally only a relatively small number of representative objects (i.e., fewer than 10) are needed to achieve adequate detection performance in an industrial setting.

Models of interaction tasks are quantitative descriptions of relationships between human temporal performance and the spatial characteristics of the interactive tasks. Examples include Fitts' law for modeling the pointing task and Accot and Zhai's steering law for the path steering task. Interaction models can be used as guidelines to design efficient user interfaces and quantitatively evaluate interaction techniques and input devices. In this paper, we introduce and experimentally verify an interaction model for a 3D object-pursuit interaction task. Object pursuit requires that a user continuously tracks an object that moves with constant velocities in a desktop virtual environment. For modeling purposes, we divide the total object-pursuit movement into a tracking phase and a correction phase. Following a two-step modeling methodology that is originally proposed in this paper, the time for the correction phase is modeled as a function of path length, path curvature, target width, and target velocity. The object-pursuit model can be used to quantitatively evaluate the efficiency of user interfaces that involve 3D interaction with moving objects.

Models of interaction tasks are quantitative descriptions of relationships between human temporal performance and the spatial characteristics of the interactive tasks. Examples include Fitts' law for modeling the pointing task and Accot and Zhai's steering law for the path steering task, etc. Models can be used as guidelines to design efficient user interfaces and quantitatively evaluate interaction techniques and input devices. In this paper, we introduce a 3D object pursuit interaction task, in which users are required to continuously track a moving target in a virtual environment. The entire movement of the task is broken into a tracking phase and a correction phase. For each phase, we propose a model that has been verified by two experiments. As the experimental results show, the time for the tracking phase is fixed once a task has been established, while the time for the correction phase usually varies according to some characteristics of the task. It can be modeled as a function of path length, target width and the velocity with which the target moves. The proposed model can be used to quantitatively evaluate the efficiency of user interfaces that involve the interaction with moving objects.

In this article, we investigate the perception of optically simulated haptic feedback. The perception of optically and mechanically simulated bumps and holes was tested experimentally. In an earlier article, we have described the active cursor technique, a method to simulate haptic feedback optically without resorting to special mechanical force feedback devices, commonly applied to produce haptic percepts in computer interfaces. The operation of the force feedback device is substituted by tiny displacements on the cursor position relative to the intended force. This method exploits the domination of the visual over the haptic modality. Results show that people can recognize optically simulated bump and hole structures and that active cursor displacements influence the haptic perception of bumps and holes. Depending on the simulated strength of the force, optically simulated haptic feedback can take precedence over mechanically simulated haptic feedback and also the other way around. When optically simulated and mechanically simulated haptic feedback counteract each other, however, the weight attributed to each source of haptic information differs from user to user. It is concluded that active cursor displacements can be used to simulate the operation of mechanical force feedback devices.

Display systems typically operate at a minimum rate of 60 Hz. However, existing VR-architectures generally produce application updates at a lower rate. Consequently, the display is not updated by the application every display frame. This causes a number of undesirable perceptual artifacts. We describe an architecture that provides a programmable display layer (PDL) in order to generate updated display frames. This replaces the default display behavior of repeating application frames until an update is available. We will show three benefits of the architecture typical to VR. First, smooth motion is provided by generating intermediate display frames by per-pixel depth-image warping using 3D motion fields. Smooth motion eliminates various perceptual artifacts due to judder. Second, we implement fine-grained latency reduction at the display frame level using a synchronized prediction of simulation objects and the viewpoint. This improves the average quality and consistency of latency reduction. Third, a crosstalk reduction algorithm for consecutive display frames is implemented, which improves the quality of stereoscopic images. To evaluate the architecture, we compare image quality and latency to that of a classic level-of-detail approach.

The study of aimed movements has a long history, starting at least as far back as 1899 when Wood-worth proposed a two-component model in which aimed movements are broken into an initial ballistic phase and an additional control phase. In this paper, we use Wood-worth's model for experimentally comparing aimed movements in the real world with those in a virtual environment. Trajectories from real world movements have been collected and compared to trajectories of movements taken from a virtual environment. From this, we show that significant temporal differences arise in both the ballistic and control phases, but the difference is much larger in the control phase; users' improvement is relatively greater in the virtual world than in the real world. They progress more in ballistic phase in the real world, but more in correction phase in the virtual world. These results allow us to better understand the pointing tasks in virtual environments.

Designing low end-to-end latency system architectures for virtual reality is still an open and challenging problem. We describe the design, implementation and evaluation of a client-server depth-image warping architecture that updates and displays the scene graph at the refresh rate of the display. Our approach works for scenes consisting of dynamic and interactive objects. The end-to-end latency is minimized as well as smooth object motion generated. However, this comes at the expense of image quality inherent to warping techniques. We evaluate the architecture and its design trade-offs by comparing latency and image quality to a conventional rendering system. Our experience with the system confirms that the approach facilitates common interaction tasks such as navigation and object manipulation.

We describe an architecture that provides a programmable display layer in order to allow the execution of custom programs on consecutive display frames. This replaces the default display behavior of repeating application frames until an update is available. The architecture is implemented using a multi-GPU system. We will show three applications of the architecture typical to VR. First, smooth motion is provided by generating intermediate display frames by per-pixel depth-image warping using 3D motion fields. Smooth motion can be beneficial for walk-throughs of large scenes. Second, we implement fine-grained latency reduction at the display frame level using a synchronized prediction of simulation objects and the viewpoint. This improves the average quality and consistency of latency reduction. Third, a crosstalk reduction algorithm for consecutive display frames is implemented, which improves the quality of stereoscopic images.

An imaging mass spectrometer is an analytical instrument that can determine the spatial distribution of chemical compounds on complex surfaces. The output of the device is a multi-spectral datacube; a three-dimensional (3D) dataset in which the xy-dimension represents the surface position and the z-dimension represents the mass spectral distribution. Analysts try to discover correlations in spectral profiles and spatial distributions inside a datacube. New technological developments allow mass spectrometric imaging on a higher spatial and spectral resolution. In this paper we present a parametric visualization technique which allows an analyst to examine spectral and spatially correlated patterns on the highest possible resolution. Principal component analysis (PCA) is used to decompose the datacube into several discriminating components. We represent these extracted features as abstract geometric shapes and use three parameters to allow for data exploration. The first parameter thresholds the spectral contribution at which an extracted component is visualized. The level of detail the shapes is controlled by a second parameter and a third parameter determines at which density-level the extracted feature is represented. This new visualization technique enables an analyst to select the most relevant spectral correlations and investigate their specific spatial distribution. With this method, less noise is included in the visualization of extracted features and by introducing various levels of detail the full spectral resolution can be utilized.

Stereo displays suffer from crosstalk, an effect that reduces or even inhibits the viewer's ability to correctly perceive depth. Previous work on software crosstalk reduction focussed on the preprocessing of static scenes which are viewed from a fixed viewpoint. However, in virtual environments scenes are dynamic, and are viewed from various viewpoints in real-time on large display areas. In this paper, three methods are introduced for reducing crosstalk in virtual environments. A non-uniform crosstalk model is described, which can be used to accurately reduce crosstalk on large display areas. In addition, a novel temporal algorithm is used to address the problems that occur when reducing crosstalk in dynamic scenes. This way, high-frequency jitter caused by the erroneous assumption of static scenes can be eliminated. Finally, a perception based metric is developed that allows us to quantify crosstalk. We provide a detailed description of the methods, discuss their tradeoffs, and compare their performance with existing crosstalk reduction methods

We introduce an architecture for smooth motion in virtual environments. The system performs forward depth image warping to produce images at video refresh rates. In addition to color and depth, our 3D warping approach records per-pixel motion information during rendering of the three-dimensional scene. These enhanced depth images are used to perform per-pixel advection, which considers object motion and view changes. Our dual graphics card architecture is able to render the 3D scene at the highest possible frame rate on one graphics card, while doing the depth image warping on a second graphics engine at video refresh rate.

The object-oriented visualization Toolkit (VTK) is widely used for scientific visualization. VTK is a visualization library that provides a large number of functions for presenting three-dimensional data. Interaction with the visualized data is controlled with two-dimensional input devices, such as mouse and keyboard. Support for real three-dimensional and multimodal input is non-existent. This paper describes VR-VTK: a multimodal interface to VTK on a virtual environment. Six degree of freedom input devices are used for spatial 3D interaction. They control the 3D widgets that are used to interact with the visualized data. Head tracking is used for camera control. Pedals are used for clutching. Speech input is used for application commands and system control. To address several problems specific for spatial 3D interaction, a number of additional features, such as more complex interaction methods and enhanced depth perception, are discussed. Furthermore, the need for multimodal input to support interaction with the visualization is shown. Two existing VTK applications are ported using VR-VTK to run in a desktop virtual reality system. Informal user experiences are presented.

Determining the shape of coral structures is an essential step for coral biologists to classify and compare corals. Currently, coral biologists analyze shape by performing manual measurements on photographs of coral colonies. In this paper we describe an interactive visualization system for measuring coral shapes in a robust and quantitative way. The input of the system is a CT scan of the coral, and the output consists of statistical distributions of various morphological properties. The approach is to first extract a skeleton from the CT scan, and then to perform measurements on the skeleton graph. Interactive visualization is necessary, since various features of the coral prevent the system from being fully automatic.

Current practice in particle visualization renders particle position data directly onto the screen as points or glyphs. Using a camera placed at a fixed position, particle motions can be visualized by rendering trajectories or by animations. Applying such direct techniques to large, time dependent particle data sets often results in cluttered images in which the dynamic properties of the underlying system are difficult to interpret. In this case study we take an alternative approach to the visualization of ion motions. Instead of rendering ion position data directly, we first extract meaningful motion information from the ion position data and then map this information onto geometric primitives. Our goal is to produce high-level visualizations that reflect the physicists' way of thinking about ion dynamics. Parameterized geometric icons are defined to encode motion information of clusters of related ions. In addition, a parameterized camera control mechanism is used to analyze relative instead of only absolute ion motions. We apply the techniques to simulations of Fourier transform mass spectrometry (FTMS) experiments. The data produced by such simulations can amount to 5 10(4) ions and 10(5) timesteps. This paper discusses the requirements, design and informal evaluation of the implemented system.

In this paper we propose an approach in which interactive visualization and analysis are combined with batch tools for the processing of large data collections. Large and heterogeneous data collections are difficult to analyze and pose specific problems to interactive visualization. Application of the traditional interactive processing and visualization approaches as well as batch processing encounter considerable drawbacks for such large and heterogeneous data collections due to the amount and type of data. Computing resources are not sufficient for interactive exploration of the data and automated analysis has the disadvantage that the user has only limited control and feedback on the analysis process. In our approach, an analysis procedure with features and attributes of interest for the analysis is defined interactively. This procedure is used for off-line processing of large collections of data sets. The results of the batch process along with "visual summaries" are used for further analysis. Visualization is not only used for the presentation of the result, but also as a tool to monitor the validity and quality of the operations performed during the batch process. Operations such as feature extraction and attribute calculation of the collected data sets are validated by visual inspection. This approach is illustrated by an extensive case study, in which a collection of confocal microscopy data sets is analyzed.

In this paper we present a new application of the principal component analysis (PCA) to generate multidimensional transfer functions. These transfer functions are needed in the volumetric visualization of spectral data to isolate regions that contain interesting peak-shaped features. Both large and small peaks can be equally important and represent the presence of different chemical elements in a dataset. Principal component analysis separates these peaks in different uncorrelated components and can simultaneously identify spatial patterns. This approach is characterized by the direct linkage between the resulting spectral and spatial components. Our method enables us to create an opacity map from these components. One or more mappings can be selected to highlight features in three-dimensional volume visualization.

To enable a user to perform Virtual Reality tasks as efficiently as possible, reducing tracking inaccuracies from noise and latency is crucial. Much work has been done to improve tracking performance by using predictive filtering methods. However, it is unclear what the benefits of each of these methods are in practice, which parameters influence their performance, and what the extent of this influence is. We present an analysis of various orientation prediction and filtering methods using various hand tasks and synthetic signals, and evaluate their performance in relation to each other. We identify critical parameters and analyse their influence on accuracy. Our results show that for the tested datasets, the use of an EKF is sufficient for orientation prediction in VR/AR.

In recent years many optical trackers have been proposed for usage in Virtual Environments. In this paper, we compare three model based optical tracking algorithms for pose determination of input devices. In particular, we study the behavior of these algorithms when applied to two-handed manipulation tasks. We experimentally show how critical parameters influence the relative accuracy, latency and robustness of each algorithm. Although the study has been performed in a specific near-field virtual environment, the results can be applied to other virtual environments such as workbenches and CAVEs.

We describe a novel optical tracker algorithm for the tracking of interaction devices in virtual and augmented reality. The tracker uses invariant properties of marker patterns to efficiently identify and reconstruct the pose of these interaction devices. Since invariant properties are sensitive to noise in the 2D marker positions, an offline training session is used to determine deviations in these properties. These deviations are taken into account when searching for the patterns once the tracker is used.

This paper introduces GraphSplatting, a technique which transforms a graph into a two-dimensional scalar field. The scalar field can be rendered as a color coded map, a height field, or a set of contours. Splat fields allow for the visualization of arbitrarily large graphs without cluttering. They provide density information which can be used to determine the structure of the graph. The construction, visualization, and interaction with splat fields is discussed. Two applications illustrate the usage of GraphSplatting.

This paper describes BM3D: a method for the analysis of motion in time dependent volume data. From a sequence of volume data sets a sequence of vector data sets representing the movement of the data is computed. A block matching technique is used for the reconstruction of data movement. The derived vector field can be used for the visualization of time dependent volume data. The method is illustrated in two applications.

In this case study we discuss an interactive feature tracking system and its use for the analysis of chromatin decondensation. Features are described as points in a multidimensional attribute space. Distances between points are used as a measure for feature correspondence. Users can interactively experiment with the correspondence measure in order to gain insight in chromatin movement. In addition, by defining time as an attribute, tracking problems related to noisy confocal data can be circumvented.

The study of time dependent characteristics of proteins is important for gaining insight into many biological processes. However, visualizing protein dynamics by animating atom trajectories does not provide satisfactory results. When the trajectory is sampled with large times steps, the impression of smooth motion will be destroyed due to the effects of temporal aliasing. Sampling with small time steps will result in the camouflage of interesting motions. In this case study, we discuss techniques for the interactive 3D visualization of the dynamics of the photoactive yellow protein. We use essential dynamics methods to filter out uninteresting atom motions from the larger concerted motions. In this way, clear and concise 3D animations of protein motions can be produced. In addition, we discuss various interactive techniques that allow exploration of the essential subspace of the protein. We discuss the merits of these techniques when applied to the analysis of the yellow protein.

The microscopic analysis of time dependent 3D live cells provides considerable challenges to visualization. Effective visualization can provide insight into the structure and functioning of living cells. The paper presents a case study in which a number of visualization techniques were applied to analyze a specific problem in cell biology: the condensation and de-condensation of chromosomes during cell division. The spatial complexity of the data required sophisticated presentation techniques. The interactive virtual reality enabled visualization system, proteus, specially equipped for time dependent 3D data sets is described. An important feature of proteus is that it is extendible to cope with application-specific demands.

Virtual environments have shown great promise as a research tool in science and engineering. In this paper, we study a classical problem in mathematics: that of approximating globally optimal Fekete point configurations. We found that a highly interactive virtual environment, combined with a time-critical computation, can provide valuable insight into the symmetry and stability of Fekete point configurations. We believe that virtual environments provide more natural interfaces to complex systems, allowing users to perceive, interpret and interact with the problem more rapidly.

Fish tank VR systems provide head coupled perspective projected stereo images on a display device of limited dimensions that resides at a fixed location. Therefore, fish tank VR systems provide only a limited virtual workspace. As a result, such systems are less suited for displaying virtual worlds that extend beyond the available workspace and depth perception problems arise when displaying objects (virtually) located on the edge of the workspace in between the viewer and the display screen. In this paper we present two techniques to reduce this disadvantage: cadre viewing and amplified head rotations. The first aims to eliminate the problems in depth perception for objects with negative parallax touching the screen surround. Subjective observations from an informal user study indicate a reduction of confusion in depth perception. The second provides a transparent navigation technique to allow users to view larger portions of the virtual world without the need for an additional input device to navigate. A user study shows it performs equally well when compared to a technique based on the use of an additional spatial input device.

Current algorithms to create depth of field (DOF) effects are either too costly to be applied in VR systems, or they produce inaccurate results. In this paper, we present a new algorithm to create DOF effects. The algorithm is based on two techniques: one of high accuracy and one of high speed but less accurate. The latter is used to create DOF effects in the peripheral viewing area where accurate results are not necessary. The first is applied to the viewing volume focussed at by the viewer. Both techniques make extensive use of rendering hardware, for texturing as well as image processing. The algorithm presented in this paper is an improvement over other (fast) DOF algorithms in that it is faster and provides better quality DOF effects where it matters most.

Visualization of topological information of a vector field can provide useful information on the structure of the field. However, in turbulent flows standard critical point visualization will result in a cluttered image which is difficult to interpret. This paper presents a technique for collapsing topologies. The governing idea is to classify the importance of the critical points in the topology. By only displaying the more important critical points, a simplified depiction of the topology can be provided. Flow consistency is maintained when collapsing the topology, resulting in a visualization which is consistent with the original topology. We apply the collapsing topology technique to a turbulent flow field.

Spot noise and line integral convolution (LIC) are two texture synthesis techniques for vector field visualization. The two techniques are compared. Continuous directional convolution is used as a common basis for comparing the techniques. It is shown that the techniques are based on the same mathematical concept. Comparisons of the visual appearance of the output and performance of the algorithms are made.

In this paper the motivation, design and application of a distributed blackboard architecture for interactive data visualization is discussed. The main advantages of the architecture are twofold. First, it allows visualization tools to be tightly integrated with simulations. Second, it allows qualitative and quantitative analysis to be combined during the visualization process.

HyperSlice is a new method for the visualization of scalar functions of many variables. With this method the multi-dimensional function is presented in a simple and easy to understand way in which all dimensions are treated identically. The central concept is the representation of a multi-dimensional function as a matrix of orthogonal two-dimensional slices. These two-dimensional slices lend themselves very well to interaction via direct manipulation, due to a one to one relation between screen space and variable space. Several interaction techniques, for navigation, the location of maxima, and the use of user-defined paths, are presented.>

Abstract This paper shows how a formal description of the GKS input model can be extended to include hierarchically structured input devices

This paper was inspired by the Components/ Frameworks approach to a Reference Model for computer graphics, currently under discussion in the ISO computer graphics subject committee. The paper shows how a formal description of the GKS input model may be given in Hoare’s CSP notation and explores some extensions in which some of the components in the GKS model are replaced by more interesting ones. The paper thus demonstrates some of the power and flexibility inherent in the Component/ Frameworks idea. The use of a formal notation led to a deepening of the authors’ understanding of the input model and suggested some different ways of looking at the input model.

Abstract The logical input device model, as is adopted in the standardized graphics package GKS, has been an accepted basis for producing device‐independent graphics systems. However, when used in highly interactive graphical applications, the logical input device model does not provide sufficient support for a number of fundamental issues inherent to interaction. This paper reopens a discussion which questions the functionality provided by the logical input device model when brought in conjunction with interaction. In particular, the logical input device model does not support the notion of input/output symmetry. CR Categories: 1.3.4 [Computer graphics]: Graphics utilities‐ graphics packages; 1.3.6 [Computer graphics]: Methodology and techniques‐ device independence, interaction techniques