C oronary CT angiography has proven prognostic value for cardiac events (1-6). It depicts vessel lumen and wall characteristics including stenoses, remodeling, plaque thickness, and degree of calcification (7). Imaging improves prognostic accuracy beyond that offered by traditional risk estimation methods, and data suggest that it might be useful for primary assessment of coronary risk under some circumstances (8-10). A practical problem has been how to score atherosclerotic features for use in prognosis estimation models. The most common approach is to divide the coronary tree into 16 segments and then score each segment according to certain simple criteria (11-13). For example, the segmental plaque score scores the amount of plaque from 0 to 3 for each segment and takes the sum. The Coronary Artery Disease Reporting and Data System (CAD-RADS), a standardized reporting system, was recently introduced for clinical use (14). These scoring systems are necessarily an abstraction from the underlying pathologic condition, and there is the chance of discarding useful information along the way. Machine learning can explore a large number of possible models and construct a good model without overlooking important input features or including unnecessary ones (15). In this study, patients were followed after coronary CT angiography for the occurrence of death and myocardial infarction. The hypothesis was that machine learning, compared with conventional scoring systems, could find a combination of arterial features that better discriminated patients who did not experience an adverse event from those who did. We analyzed data as summarized by the reading radiologists (ie, from human visual analysis, not

No abstract

No abstract

A space-resolving flux detector (SRFD) is developed to measure the X-ray flux emitted from a specified region in hohlraum with a high resolution up to 0.11mm for the first time. This novel detector has been used successfully to measure the distinct X-ray fluxes emitted from hot laser spot and cooler re-emitting region simultaneously, in the hohlraum experiments on SGIII prototype laser facility. According to our experiments, the ratio of laser spot flux to re-emitted flux shows a strong time-dependent behavior, and the area-weighted flux post-processed from the measured laser spot flux and re-emitting wall flux agrees with that measured from Laser Entrance Hole by using flat-response X-ray detector (F-XRD). The experimental observations is reestablished by our two-dimensional hydrodynamic simulations and is well understood with the power balance relationship.

No abstract

Abstract.A novel preoperative surgery planning method is proposed for percutaneous hepatic microwave ablation. An iterative framework for necrosis field simulation and 3D necrosis zone reconstruction is introduced here, and the necrosis model is further superimposed to patient anatomy structures using advanced GPU-accelerated visualization techniques. The full surgery planning is performed by the surgeon in an interactively way, until the optimal surgery plan is achieved. Experiments have been performed on realistic patient with hepatic cancer and the actual necrosis zone are measured in postoperative CT images for patient. Results show that this method is relative accurate for preoperative trajectory plan and could be used as an assistant to the clinical practice.

Measurements of iron-plasma absorption spectrum over 150–1200 eV photon energy range were reported at temperature T = (72 ± 4) eV. The electron temperature was diagnosed with the absorption spectrum of aluminum mixed with iron. The density was not diagnosed directly but obtained from a radiative hydrodynamic simulation with the Multi-1D code. The broad photon energy range enables simultaneous observation of the L-shell and M-shell transitions that dominate the radiation transport at this temperature. The spectrally resolved transmission data were compared to the detailed-configuration-accounting model calculations and reasonable agreement was found.

Purpose The purpose of this paper is to present a detection method based on computer vision for automatic flexible printed circuit (FPC) defect detection. Design/methodology/approach This paper proposes a new method of watershed segmentation based on morphology. A dimensional increment matrix calculation method and an image segmentation method combined with a fuzzy clustering algorithm are provided. The visibility of the segmented image and the segmentation accuracy of a defective image are guaranteed. Findings Compared with the traditional one, the segmentation result obtained in this study is superior in aspects of noise control and defect segmentation. It completely proves that the segmentation method proposed in this study is better matches the requirements of FPC defect extraction and can more effectively provide the segmentation result. Compared with traditional human operators, this system ensures greater accuracy and more objective detection results. Research limitations/implications The extraction of FPC defect characteristics contains some obvious characteristics as well as many implied characteristics. These characteristics can be extracted through specific space conversion and arithmetical operation. Therefore, more images are required for analysis and foresight to establish a more widely used FPC defect detection sorting algorithm. Originality/value This paper proposes a new method of watershed segmentation based on morphology. It combines a traditional edge detection algorithm and mathematical morphology. The FPC surface defect detection system can meet the requirements of online detection through constant design and improvement. Therefore, human operators will be replaced by machine vision, which can preferably reduce the production costs and improve the efficiency of FPC production.