Assessment and treatment of visuospatial neglect using active learning with Gaussian processes regression.

Visuospatial neglect is a disorder characterised by impaired awareness for visual stimuli located in regions of space and frames of reference. It is often associated with stroke. Patients can struggle with all aspects of daily living and community participation. Assessment methods are limited and show several shortcomings, considering they are mainly performed on paper and do not implement the complexity of daily life. Similarly, treatment options are sparse and often show only small improvements. We present an artificial intelligence solution designed to accurately assess a patient's visuospatial neglect in a three-dimensional setting. We implement an active learning method based on Gaussian process regression to reduce the effort it takes a patient to undergo an assessment. Furthermore, we describe how this model can be utilised in patient oriented treatment and how this opens the way to gamification, tele-rehabilitation and personalised healthcare, providing a promising avenue for improving patient engagement and rehabilitation outcomes. To validate our assessment module, we conducted clinical trials involving patients in a real-world setting. We compared the results obtained using our AI-based assessment with the widely used conventional visuospatial neglect tests currently employed in clinical practice. The validation process serves to establish the accuracy and reliability of our model, confirming its potential as a valuable tool for diagnosing and monitoring visuospatial neglect. Our VR application proves to be more sensitive, while intra-rater reliability remains high.

On the Angular Control of Rotating Lasers by Means of Line Calculus on Hyperboloids.

We propose a new paradigm for modelling and calibrating laser scanners with rotation symmetry, as is the case for lidars or for galvanometric laser systems with one or two rotating mirrors. Instead of bothering about the intrinsic parameters of a physical model, we use the geometric properties of the device to model it as a specific configuration of lines, which can be recovered by a line-data-driven procedure. Compared to universal data-driven methods that train general line models, our algebraic-geometric approach only requires a few measurements. We elaborate the case of a galvanometric laser scanner with two mirrors, that we model as a grid of hyperboloids represented by a grid of 3×3 lines. This provides a new type of look-up table, containing not more than nine elements, lines rather than points, where we replace the approximating interpolation with exact affine combinations of lines. The proposed method is validated in a realistic virtual setting. As a collateral contribution, we present a robust algorithm for fitting ruled surfaces of revolution on noisy line measurements.

Accuracy Assessment of Joint Angles Estimated from 2D and 3D Camera Measurements.

To automatically evaluate the ergonomics of workers, 3D skeletons are needed. Most ergonomic assessment methods, like REBA, are based on the different 3D joint angles. Thanks to the huge amount of training data, 2D skeleton detectors have become very accurate. In this work, we test three methods to calculate 3D skeletons from 2D detections: using the depth from a single RealSense range camera, triangulating the joints using multiple cameras, and combining the triangulation of multiple camera pairs. We tested the methods using recordings of a person doing different assembly tasks. We compared the resulting joint angles to the ground truth of a VICON marker-based tracking system. The resulting RMS angle error for the triangulation methods is between 12° and 16°, showing that they are accurate enough to calculate a useful ergonomic score from.

Extrinsic Camera Calibration with Line-Laser Projection.

Knowledge of precise camera poses is vital for multi-camera setups. Camera intrinsics can be obtained for each camera separately in lab conditions. For fixed multi-camera setups, the extrinsic calibration can only be done in situ. Usually, some markers are used, like checkerboards, requiring some level of overlap between cameras. In this work, we propose a method for cases with little or no overlap. Laser lines are projected on a plane (e.g., floor or wall) using a laser line projector. The pose of the plane and cameras is then optimized using bundle adjustment to match the lines seen by the cameras. To find the extrinsic calibration, only a partial overlap between the laser lines and the field of view of the cameras is needed. Real-world experiments were conducted both with and without overlapping fields of view, resulting in rotation errors below 0.5°. We show that the accuracy is comparable to other state-of-the-art methods while offering a more practical procedure. The method can also be used in large-scale applications and can be fully automated.

Interactive Camera Network Design Using a Virtual Reality Interface.

The traditional literature on camera network design focuses on constructing automated algorithms. These require problem-specific input from experts in order to produce their output. The nature of the required input is highly unintuitive, leading to an impractical workflow for human operators. In this work we focus on developing a virtual reality user interface allowing human operators to manually design camera networks in an intuitive manner. From real world practical examples we conclude that the camera networks designed using this interface are highly competitive with, or sometimes even superior to, those generated by automated algorithms, but the associated workflow is more intuitive and simple. The competitiveness of the human-generated camera networks is remarkable because the structure of the optimization problem is a well known combinatorial NP-hard problem. These results indicate that human operators can be used in challenging geometrical combinatorial optimization problems, given an intuitive visualization of the problem.

Camera Calibration Using Gray Code.

In order to determine camera parameters, a calibration procedure involving the camera recordings of a checkerboard is usually performed. In this paper, we propose an alternative approach that uses Gray-code patterns displayed on an LCD screen. Gray-code patterns allow us to decode 3D location information of points of the LCD screen at every pixel in the camera image. This is in contrast to checkerboard patterns where the number of corresponding locations is limited to the number of checkerboard corners. We show that, for the case of a UEye CMOS camera, the precision of focal-length estimation is 1.5 times more precise than when using a standard calibration with a checkerboard pattern.

Realistic Indoor Radio Propagation for Sub-GHz Communication.

This research article proposes a novel ray-launching propagation loss model that is able to use an environment model that contains the real geometry. This environment model is made by applying a Simultaneous Localization and Mapping (SLAM) algorithm. As a solution to the rising demands of Internet of Things applications for indoor environments, this deterministic radio propagation loss model is able to simulate an accurate coverage map that can be used for localization applications or network optimizations. Since this propagation loss model uses a 2D environment model that was captured by a moving robot, an automated validation model is developed so that a wireless sensor network can be used for validating the propagation loss model. We validated the propagation loss model by evaluated two environment models towards the lowest Root Mean Square Error (RMSE), the Mean Absolute Error (MAE), and the Mean Error (ME). Furthermore, the correlation between the number of rays and the RMSE is analyzed and the correlation between the number of reflections versus the RMSE is also analyzed. Finally, the performance of the radio propagation loss model is analyzed.

Extrinsic Calibration of a Laser Galvanometric Setup and a Range Camera.

Currently, galvanometric scanning systems (like the one used in a scanning laser Doppler vibrometer) rely on a planar calibration procedure between a two-dimensional (2D) camera and the laser galvanometric scanning system to automatically aim a laser beam at a particular point on an object. In the case of nonplanar or moving objects, this calibration is not sufficiently accurate anymore. In this work, a three-dimensional (3D) calibration procedure that uses a 3D range sensor is proposed. The 3D calibration is valid for all types of objects and retains its accuracy when objects are moved between subsequent measurement campaigns. The proposed 3D calibration uses a Non-Perspective-n-Point (NPnP) problem solution. The 3D range sensor is used to calculate the position of the object under test relative to the laser galvanometric system. With this extrinsic calibration, the laser galvanometric scanning system can automatically aim a laser beam to this object. In experiments, the mean accuracy of aiming the laser beam on an object is below 10 mm for 95% of the measurements. This achieved accuracy is mainly determined by the accuracy and resolution of the 3D range sensor. The new calibration method is significantly better than the original 2D calibration method, which in our setup achieves errors below 68 mm for 95% of the measurements.

Computer-aided analysis of star shot films for high-accuracy radiation therapy treatment units.

As mechanical stability of radiation therapy treatment devices has gone beyond sub-millimeter levels, there is a rising demand for simple yet highly accurate measurement techniques to support the routine quality control of these devices. A combination of using high-resolution radiosensitive film and computer-aided analysis could provide an answer. One generally known technique is the acquisition of star shot films to determine the mechanical stability of rotations of gantries and the therapeutic beam. With computer-aided analysis, mechanical performance can be quantified as a radiation isocenter radius size. In this work, computer-aided analysis of star shot film is further refined by applying an analytical solution for the smallest intersecting circle problem, in contrast to the gradient optimization approaches used until today. An algorithm is presented and subjected to a performance test using two different types of radiosensitive film, the Kodak EDR2 radiographic film and the ISP EBT2 radiochromic film. Artificial star shots with a priori known radiation isocenter size are used to determine the systematic errors introduced by the digitization of the film and the computer analysis. The estimated uncertainty on the isocenter size measurement with the presented technique was 0.04 mm (2σ) and 0.06 mm (2σ) for radiographic and radiochromic films, respectively. As an application of the technique, a study was conducted to compare the mechanical stability of O-ring gantry systems with C-arm-based gantries. In total ten systems of five different institutions were included in this study and star shots were acquired for gantry, collimator, ring, couch rotations and gantry wobble. It was not possible to draw general conclusions about differences in mechanical performance between O-ring and C-arm gantry systems, mainly due to differences in the beam-MLC alignment procedure accuracy. Nevertheless, the best performing O-ring system in this study, a BrainLab/MHI Vero system, and the best performing C-arm system, a Varian Truebeam system, showed comparable mechanical performance: gantry isocenter radius of 0.12 and 0.09 mm, respectively, ring/couch rotation of below 0.10 mm for both systems and a wobble of 0.06 and 0.18 mm, respectively. The methodology described in this work can be used to monitor mechanical performance constancy of high-accuracy treatment devices, with means available in a clinical radiation therapy environment.