Michael Grass

Bio: Michael Grass is an academic researcher from Philips. The author has contributed to research in topics: Iterative reconstruction & Projection (set theory). The author has an hindex of 31, co-authored 299 publications receiving 4602 citations.

Comparison of Deep Learning Approaches for Multi-Label Chest X-Ray Classification.

Abstract: The increased availability of labeled X-ray image archives (e.g. ChestX-ray14 dataset) has triggered a growing interest in deep learning techniques. To provide better insight into the different approaches, and their applications to chest X-ray classification, we investigate a powerful network architecture in detail: the ResNet-50. Building on prior work in this domain, we consider transfer learning with and without fine-tuning as well as the training of a dedicated X-ray network from scratch. To leverage the high spatial resolution of X-ray data, we also include an extended ResNet-50 architecture, and a network integrating non-image data (patient age, gender and acquisition type) in the classification process. In a concluding experiment, we also investigate multiple ResNet depths (i.e. ResNet-38 and ResNet-101). In a systematic evaluation, using 5-fold re-sampling and a multi-label loss function, we compare the performance of the different approaches for pathology classification by ROC statistics and analyze differences between the classifiers using rank correlation. Overall, we observe a considerable spread in the achieved performance and conclude that the X-ray-specific ResNet-38, integrating non-image data yields the best overall results. Furthermore, class activation maps are used to understand the classification process, and a detailed analysis of the impact of non-image features is provided.

Noninvasive Coronary Angiography with 16–Detector Row CT: Effect of Heart Rate

Abstract: PURPOSE: To evaluate the effect of heart rate on the quality of coronary angiograms obtained with 16–detector row computed tomography (CT) by using temporally enhanced three-dimensional (3D) approaches. MATERIALS AND METHODS: The local ethics committee approved the study, and informed consent was obtained from all patients. Fifty patients underwent coronary CT angiography (heart rate range, 45–103 beats per minute). Raw data from helical CT and electrocardiography (ECG) were saved in a combined data set. Retrospectively ECG-gated images were reconstructed at preselected phases (50% and 80%) of the cardiac cycle. A 3D voxel-based approach with cardiac phase weighting was used for reconstruction. Testing for correlation between heart rate, cardiac phase reconstruction window, and image quality was performed with Kruskal-Wallis analysis. Image quality (freedom from cardiac motion–related artifacts) was referenced against findings at conventional angiography in a secondary evaluation step. Regression analysis...

3D cone-beam CT reconstruction for circular trajectories

Abstract: 3D reconstruction from 2D projections obtained along a single circular source trajectory is most commonly done using an algorithm due to Feldkamp, Davis and Kress. In this paper we propose an alternative approach based on a cone-beam to parallel-beam rebinning step, a corresponding rebinning step into a rectangular virtual detector plane and a filtered backprojection. This approach yields an improved image quality reflected by a decreased low-intensity drop which is well known for 3D reconstruction from projection data obtained along circular trajectories. At the same time the computational complexity is lower than in Feldkamp's original approach. Based on this idea, a hybrid 3D cone-beam reconstruction method is formulated that enlarges the reconstruction volume in its dimension along the rotation axis of the cone-beam CT system. This enlargement is achieved by applying different reconstruction conditions for each voxel. An optimal ratio between the reconstructible and irradiated volume of the scanned object is achieved.

Motion-compensated and gated cone beam filtered back-projection for 3-D rotational X-ray angiography

Abstract: This paper presents a method to reconstruct moving objects from cone beam X-ray projections acquired during a single rotational run using a given motion vector field. The method is applicable to voxel driven cone-beam filtered back-projection reconstruction approaches. Here, a formulation based on the algorithm of Feldkamp, Davis, and Kress (FDK) is presented. The motion correction is applied during the back-projection step by shifting the voxel to be reconstructed according to the motion vector field. The method is applied to three-dimensional (3-D) rotational X-ray angiography. Projections from a beating coronary heart phantom are simulated. Motion-compensated reconstructions with varying accuracy of the applied motion field are carried out for a late diastolic heart phase and compared to the reconstruction obtained with the standard FDK-method from projections of the corresponding motion-free model in the same heart phase. Furthermore, gated reconstructions are calculated by weighting the projections according to their cardiac phase without using a motion vector field. Different gating window widths are applied, and the reconstructions are compared. Using the correct motion field with the motion-compensated reconstruction, the image quality of the standard reconstruction from the corresponding motion-free coronary model can almost be recovered. The reconstructed image quality stays acceptable if the accuracy of the motion field sampling points is better than 1 mm. The gated reconstructions with a window width of 15%-20% of the cardiac cycle lead to superior results compared to nearest neighbor gating, especially for histogram based visualization and analysis. The motion-compensated reconstructions provide sharp images of the coronaries far surpassing the image quality of gated reconstructions

Assessment of coronary artery disease by cardiac computed tomography : A scientific statement from the american heart association committee on cardiovascular imaging and intervention, council on cardiovascular radiology and intervention, and committee on cardiac imaging, council on clinical cardiology

Abstract: This scientific statement reviews the scientific data for cardiac computed tomography (CT) related to imaging of coronary artery disease (CAD) and atherosclerosis. Cardiac CT is a CT imaging technique that accounts for cardiac motion, typically through the use of ECG gating. The utility and limitations of generations of cardiac CT systems are reviewed in this statement with emphasis on CT measurement of CAD and coronary artery calcified plaque (CACP) and noncalcified plaque. Successive generations of CT technology have been applied to cardiac imaging beginning in the early 1980s with conventional CT, electron beam CT (EBCT) in 1987, and multidetector CT (MDCT) in 1999. Compared with other imaging modalities, cardiac CT has undergone an accelerated …

A review of vessel extraction techniques and algorithms

Abstract: Vessel segmentation algorithms are the critical components of circulatory blood vessel analysis systems. We present a survey of vessel extraction techniques and algorithms. We put the various vessel extraction approaches and techniques in perspective by means of a classification of the existing research. While we have mainly targeted the extraction of blood vessels, neurosvascular structure in particular, we have also reviewed some of the segmentation methods for the tubular objects that show similar characteristics to vessels. We have divided vessel segmentation algorithms and techniques into six main categories: (1) pattern recognition techniques, (2) model-based approaches, (3) tracking-based approaches, (4) artificial intelligence-based approaches, (5) neural network-based approaches, and (6) tube-like object detection approaches. Some of these categories are further divided into subcategories. We have also created tables to compare the papers in each category against such criteria as dimensionality, input type, preprocessing, user interaction, and result type.

A three-dimensional statistical approach to improved image quality for multislice helical CT.

Abstract: Multislice helical computed tomography scanning offers the advantages of faster acquisition and wide organ coverage for routine clinical diagnostic purposes. However, image reconstruction is faced with the challenges of three-dimensional cone-beam geometry, data completeness issues, and low dosage. Of all available reconstruction methods, statistical iterative reconstruction (IR) techniques appear particularly promising since they provide the flexibility of accurate physical noise modeling and geometric system description. In this paper, we present the application of Bayesian iterative algorithms to real 3D multislice helical data to demonstrate significant image quality improvement over conventional techniques. We also introduce a novel prior distribution designed to provide flexibility in its parameters to fine-tune image quality. Specifically, enhanced image resolution and lower noise have been achieved, concurrently with the reduction of helical cone-beam artifacts, as demonstrated by phantom studies. Clinical results also illustrate the capabilities of the algorithm on real patient data. Although computational load remains a significant challenge for practical development, superior image quality combined with advancements in computing technology make IR techniques a legitimate candidate for future clinical applications.