Neonatal chest computed radiography: image processing and optimal image display

OBJECTIVE: The purpose of this study was to determine soft-copy image display preferences of brightness, latitude, and detail contrast for neonatal chest computed radiography to establish a baseline for future work on low-dose imaging. CONCLUSION: Observers preferred brighter images with higher detail contrast and narrow to middle latitude for soft-copy display compared with the typical screen-film hard-copy appearance. Future research on low-dose neonatal chest imaging will be facilitated by an understanding of optimal soft-copy image display.     

On the work of the radiologist--separation of image capture from image display

A primary function of the radiologist is that of extracting sufficient information from images, in a finite time, so as to arrive at a patient management decision. Clinically important information must be present in the images and this information must be correctly perceived. This essay discusses the work of the diagnostician viewing prerecorded images. The potentials for enhancing this work by means of different image display techniques are examined. The application of stimulable phosphor computed radiographic techniques are discussed in detail as an example of a technology which clearly separates crucial portions of the imaging process. This separation of functions might improve general radiology in the near future. The principals of the discussion are not specific to any imaging modality or to any individual display technology.     

Ventilation and perfusion display in a single image

A new method of ventilation and perfusion display onto a single image is presented. From the data on regions of interest of the lungs, three-dimensional histograms are created, containing as parameters X and Y for the position of the pixels, Z for the perfusion and colour for local ventilation. The perfusion value is supplied by sets of curves having Z proportional to the local perfusion count rate. Ventilation modulates colour. Four perspective views of the histogram are simultaneously displayed to allow visualization of the entire organ. Information about the normal ranges for both ventilation and perfusion is also provided in the histograms.     

PACS and Web-based image distribution and display.

Picture archiving and communication system (PACS) delivers images to the display workstations mostly through digital image communication in medicine (DICOM) protocols in radiology departments, and there are lots of medical applications in healthcare community needing to access PACS images for different application purposes. In this paper, we first reviewed a hospital-integrated PACS image data flow and typical diagnostic display software architecture, and discussed some Web technologies and Web-based image application server architectures, as well as image accessing and viewing methods in these architectures. Then, we present one approach to develop component-based image display architecture and use image processing and display component to build a diagnostic display workstation, and also, give a method to integrate this component into Web-based image distribution server to enable users using Web browsers to access, view and manipulate PACS DICOM images as easy as with PACS display workstations. Finally, we test and evaluate the performance of image loading and displaying by using the diagnostic display workstation and the component-based Web display system, the experimental results show that the image distribution and display performance from the Web server to browser clients is similar with that of the image loading and displaying procedure of the diagnostic workstation as more browser clients accessing the Web server at same time. We also discuss the advantages and disadvantages of the Web-based image distribution and display in different medical applications.     

Ventilation and perfusion display in a single image

A new method of ventilation and perfusion display onto a single image is presented. From the data on regions of interest of the lungs, three-dimensional histograms are created, containing as parametersX andY for the position of the pixels,Z for the perfusion and colour for local ventilation. The perfusion value is supplied by sets of curves havingZ proportional to the local perfusion count rate. Ventilation modulates colour. Four perspective views of the histogram are simultaneously displayed to allow visualization of the entire organ. Information about the normal ranges for both ventilation and perfusion is also provided in the histograms.     

Benefits of helmet-mounted display image stabilisation under whole-body vibration

The effects of whole-body vertical vibration in the range 2.5-25 Hz on visual performance with two types of raster scan helmet-mounted display have been determined. The benefit of an image stabilisation system on numeral reading performance during vibration was also assessed with both display systems. Increases in mean reading time of over 130%/m . s-2 R.M.S. and increases in percentage reading error of more than 30%/m . s-2 R.M.S. were recorded with unstabilised displays. With vertical and horizontal image stabilisation, these decrements in performance were reduced to less than 40%/m . s-2 R.M.S. increase in reading time and less than 10%/m . s-2 R.M.S. increase in reading error. Data on the transmission of vibration from the seat to the head and from the head to the helmet were also obtained. These indicate a relation between biodynamic behaviour and visual performance during vibration.     

Flight trial of a helmet-mounted display image stabilisation system

An image stabilisation system for improving reading performance with a helmet-mounted display (HMD) during whole-body vibration was tested at night in a helicopter. Six subjects read arrays of 50 numerals as quickly and as accurately as possible while flying in three different flight conditions. The mean reading time for each array while stationary on the ground was approximately 21 s, and the mean reading error was 0.4% without stabilisation. In-flight mean reading time increased to approximately 40 s, and reading error was 18% without the stabilisation system. Stabilising the image significantly reduced the mean in-flight reading time to approximately 25 s with a 4% reading error. Data from the flight trial support the results of previous experiments, which suggest that HMD reading performance with vibration and night viewing conditions may be inferior to performance with daylight conditions.     

The role of the translation table in cardiac image display

Summary  No translation table is perfect. Some tables can assist in the identification of defects by highlighting contrast in the image, with either different color scales or gammas. However, this runs the risk of decreasing specificity and increasing the number of false-positive findings. Tables like this will tend to emphasize defects at certain points on the table, usually where there is a color change, and will create the perception of a change at that point, whereas changes of similar magnitude at other points will be much less apparent. A purist would make the argument that the reader, not the image display parameters, should be what determines whether a study is normal or abnormal and would argue for the use of a linear gray scale, the translation table that most faithfully reflects the uptake on different parts of the image. However, there is no one correct table to use; rather, it is more a matter of preference and familiarity. The reader must be aware that the image will appear differently on different translation tables and with color scales and should proceed cautiously when viewing an image on an unfamiliar translation table or with unfamiliar medium. These differences strongly support the case for using image quantitation, because it is the actual numbers, representing counts of radiotracer distribution, that constitute the image and are the most objective determinant as to whether a defect is real or not. If the reader is aware of this and is able to incorporate quantitation in the final interpretation, then he or she is more likely to make the correct interpretation.     

Digital image management: networking, display, and archiving

The requirements for implementing a radiology imaging network are similar to those for local area networks now being designed for other purposes to manage large data films. A radiology department serving a 500-bed hospital generates about 927 megabytes of digitally formatted data per working day. These data are expected to be on line for the patient's hospitalization period. The retrieval rate of these data among the interactive diagnosis display stations requires data throughput rates of between 2 and 5 megabits per second. This throughput rate requires signaling rates of between 20 and 50 megabits per second. Analog hard-copy generation of the images on the network is required by the referring physician for selected images that support the consultation report. Digital laser recorders using paper may be quite satisfactory. Long-term archiving must be low in cost and requires a database scheme capable of managing more than a terabyte of image data. Radiology networks must be required to bridge with other hospital information systems.     

颅内囊状未破裂动脉瘤MRI影像标注专家共识

人工智能(AI)技术的不断发展使得影像数据分析更加精准,由此对大数据的规范化使用提出了更高要求。准确标注是实现数据信息价值最大化的重要前提,直接影响AI模型的应用效能。目前AI技术已在颅内动脉瘤检出及破裂风险评估等方面开展研究。为了规范颅内动脉瘤MRI数据的标注操作及处理流程,中国女医师协会医学影像专委会及中国脑动脉瘤智能影像诊治评估协作组联合人工智能产学研领域的专家,经过反复讨论,提出颅内囊状未破裂动脉瘤的MRI标注初步指导意见,为建设该病的影像数据库提供有力保障,从而推动AI在临床的实际应用。     

Psychophysical analysis of monitor display functions affecting observer diagnostic performance of CT image on liquid crystal display monitors

The aim of the present study was to propose suitable display functions for CT image representation on liquid crystal display (LCD) monitors by analyzing the characteristics of the monitors typical display functions using psychophysical analysis. The luminance of the LCD monitor was adjusted to a maximum of 275 cd/m² and 480 cd/m². Three types of postcalibrated display functions (i.e., GSDF, CIELAB, and Exponential 2.2) were evaluated. Luminance calculation of a new grayscale test pattern (NGTP) was done for the conversion of the digital driving level (DDL) into the CT value. The psychophysical gradient of display functions for the CT value was evaluated and compared via statistical analysis. The value of GSDF and CIE decreased exponentially; however, the value of Exponential 2.2 showed a convex curve with a peak at a specific point. There was a statistically significant difference among the values of the three types of display functions on the 480 cd/m² maximum via Kruskal Wallis test (P<0.001). The GSDF was suitable for observation of abdominal and lung CT images; however, the display function combined the Exponential 2.2 and the GSDF functions and was ideal for observation of brain CT images by psychophysical analysis.     

Multi-modality image combination: five techniques for simultaneous MR-SPECT display.

Increasingly, images are being acquired of the same patient using two or more diagnostic imaging modalities. If one such modality is MRI then a data set showing essentially anatomical information is produced. If the second is SPECT using HMPAO then a data set showing cerebral perfusion is also produced. The ability to conjoin such anatomical and functional data is an important goal in radiology. In this technical report five display strategies are investigated as a means of conveying simultaneously to the radiologist all the information from two such data sets. Illustrative examples are given for each display technique. Preliminary observations are made regarding comprehensibility, information loss and efficiency in conveying all the information simultaneously, for all five techniques.