深度学习在医学影像中的应用综述

深度学习能自动从大样本数据中学习获得优良的特征表达,有效提升各种机器学习任务的性能,已广泛应用于信号处理、计算机视觉和自然语言处理等诸多领域。基于深度学习的医学影像智能计算是目前智慧医疗领域的研究热点,其中深度学习方法已经应用于医学影像处理、分析的全流程。由于医学影像内在的特殊性、复杂性,特别是考虑到医学影像领域普遍存在的小样本问题,相关学习任务和应用场景对深度学习方法提出了新要求。本文以临床常用的X射线、超声、计算机断层扫描和磁共振等4种影像为例,对深度学习在医学影像中的应用现状进行综述,特别面向图像重建、病灶检测、图像分割、图像配准和计算机辅助诊断这5大任务的主要深度学习方法的进展进行介绍,并对发展趋势进行展望。     

Development and evaluation of a context-based document representation for searching the medical literature

Received: 12 March 1997 / Revised: 26 June 1997 / Accepted: 5 August 1997     

Supporting digitally enhanced learning through measurement in higher education: Development and validation of a university students' digital competence scale

Digital competence is critical for university students to adapt to and benefit from digitally enhanced learning. Prior studies on its measurement mostly focus on educators and relied on factor analyses. However, there is a lack of valid and convenient tools to measure university students' digital competence. This study aimed to develop a digital competence scale for university students (DC-US) in digitally enhanced learning with robust psychometric properties. An initial DC-US with 23 items was proposed to measure the single latent trait of digital competence. It was validated and refined continuously through a pilot study, a main study and a predictive validity study in three datasets involving 825 participants altogether, using factor analyses, Rasch analyses and the partial least squares modelling. The final DC-US turned out to comprise two subscales: technical literacy and digital skills, with 10 items retained, and manifested high internal consistency, unidimensionality and measurement invariance. The scale also demonstrated strong predictive validity, with technical literacy greatly predicting digital skills, which negatively predicted technostress. The DC-US enables instructors and school administrators to conveniently obtain preliminary information of university students' digital competence, informing their digital class preparation and development of timely interventions for addressing digital deficiencies.     

Application of digitally processed and enhanced LANDSAT imagery for geological mapping and mineral targeting in the Singhbhum Precambrian mineralized belt,Bihar-Orissa

Standard methods of digital processing and enhancement techniques have been followed to generate linearly stretched band products, band-to-band ratio images and normal and ratio colour composites from LANDSAT multispectral data on computer compatible tapes for nine test areas in the Singhbhum Precambrian mineralized belt. The software has been developed at the Space Applications Centre Laboratory, Ahmedabad, The data products have been subjected to general assessment and geological evaluation by visual interpretation methods. The study has resulted in a number of significant conclusions regarding utility and optimization of data products, stratigraphic lithologic and structural interpretations, identifying guides for mineralization and spectral characterization of major rock types.     

一款高精度数控振荡器设计与实现

数控振荡器是全数字锁相环的关键部件,为其提供高频输出时钟。数控振荡器的性能直接影响全数字锁相环的频率范围和抖动性能。提出了一种基于全数字标准单元库设计的数控振荡器,该结构采用粗调、中调和精调级联的调节机制,实现了0.5GHz~2.6GHz的高频率范围和0.8 ps的高调节精度。在先进工艺下实现了该数控振荡器设计,并基于此数控振荡器完成了全数字锁相环的系统设计,系统抖动小于2 ps,功耗10 mW。     

Towards a better understanding of annotation tools for medical imaging: a survey

Multimedia Tools and Applications - Medical imaging refers to several different technologies that are used to view the human body to diagnose, monitor, or treat medical conditions. It requires...     

医学图像描述综述：编码、解码及最新进展

随着医疗成像技术的不断提升,放射科医师每天要撰写的医学报告也与日俱增。深度学习兴起后,基于深度学习的医学图像描述技术用于自动生成医学报告,取得了显著效果。本文全面整理了近年来深度医学图像描述方向的论文,包括这一领域的最新方法、数据集和评价指标,分析了它们各自的优劣,并以模型结构为线索予以介绍,是国内首篇针对医疗图像描述任务的综述。现今的深度医疗图像描述技术主要以编码器—解码器结构为基础进行拓展,包括但不局限于加入检索方法、模板匹配方法、注意力机制、强化学习和知识图谱等方法。检索和模板匹配方法虽然简单,但由于医学报告的特殊性仍在本任务上有不错的效果;注意力机制使模型产生报告时能关注图像和文本的某一部分,已经被几乎所有主流模型所采用;强化学习方法突破了医疗图像描述任务中梯度下降训练法与离散的语言生成评价指标不匹配的瓶颈;知识图谱方法则融合了人类医生对于疾病的先验知识,有效提高了生成报告的临床准确性。此外,Transformer等新型结构也正越来越多地取代循环神经网络（recurrent neural network,RNN）甚至卷积神经网络（convolutional neural network,CNN）的位置成为网络主干。本文最后讨论了目前深度医疗图像描述仍需解决的问题以及未来的研究方向,希望能推动深度医疗图像描述技术真正落地。     

Considering something 'ELSE': ethical, legal and socio-economic factors in medical imaging and medical informatics

The focus on the use of existing and new technologies to facilitate advances in medical imaging and medical informatics (MIMI) is often directed to the technical capabilities and possibilities that these technologies bring. The technologies, though, in acting as a mediating agent alter the dynamics and context of information delivery in subtle ways. While these changes bring benefits in more efficient information transfer and offer the potential of better healthcare, they also disrupt traditional processes and practices which have been formulated for a different setting. The governance processes that underpin core ethical principles, such as patient confidentiality and informed consent, may no longer be appropriate in a new technological context. Therefore, in addition to discussing new methodologies, techniques and applications, there is need for a discussion of ethical, legal and socio-economic (ELSE) issues surrounding the use and application of technologies in MIMI. Consideration of these issues is especially important for the area of medical informatics which after all exists to support patients, healthcare practitioners and inform science. This paper brings to light some important ethical, legal and socio-economic issues related to MIMI with the aim of furthering an interdisciplinary approach to the increasing use of Information and Communication Technologies (ICT) in healthcare.     

The European Community: standardization in medical informatics and imaging.

The use of informatics and telecommunications in health care and medicine has reached a stage where the application of standards is an absolute requirement. The AIM Programme of the CEC DGXIII stimulated the study of the subject, and has supported the set up of an official European platform for standardization in this field, the CEN TC251. The needs for standardization in one of the subareas, medical imaging, are described in detail. The future solutions of the problems concerned are regarded prerequisites for the general and practical use of PACS and IMACS.     

一种宽带数控模拟复相关器的设计及实现

介绍了一种加法型增益与偏置可数字控制的1.5~2.5GHz模拟复相关器的设计原理及实现过程,并分别在点频与宽带输入信号情况下,评估了复相关器的等效相关带宽与相位的测量精度,以及在不同输入功率情况下的信噪比。实验结果表明,此模拟复相关器的增益与偏置可实现数字自动调整,在1GHz工作带宽内幅度变化不超过1.5dB,等效相关噪声带宽达到0.905GHz,相位测量精度优于2.5°,在输入功率为-13dBm时,信噪比达到13dB。     

Robot vision: An evaluation of imaging sensors

Robot vision, robotics vision, and computer vision are terms that have evolved over the past few years to include two separate but related functions that provide visual sensing for computerdriven robots. These two functions are “electro-optical imaging” and “image processing.” The purpose of electro-optical imaging is to convert optical radiation to an appropriate electronic signal for input to the robot's computer; whereas, the purpose of image processing is to extract useful information from the electronic image provided by the sensor. This article deals with the electro-optical sensing part of robot vision. It describes the operation and properties of electro-optical imaging sensors. The fundamental concepts of optical radiation, optical radiation quantities and units, and photon energies are defined. The fundamental principles for detecting optical radiation and definitions for the primary performance measures for optical detectors are given. A relatively comprehensive overview of the parameters used to describe imaging sensors is presented. Factors affecting these parameters and obtainable sensor performance are compared. Example design analyses are presented to show the interaction of various performance measures.     

Development and evaluation of a Venus flytrap-inspired microrobot

Nature has provided the inspiration for many robots, leading to the development of biomimetic machines based on stick insects, jellyfish, butterflies, lobsters, and inchworms. Some carnivorous plants are capable of rapid motion, including mimosa, Venus flytraps, telegraph plants, sundews, and bladderworts, all of which are of interest in the design of biomimetic robots that can be activated in a controlled manner to capture prey using trigger hairs. Here, we describe a biomimetic robotic inspired by a Venus flytrap and fabricated using two ionic polymer metal composite (IPMC) actuators. First, we describe the structure of the robotic flytrap, which consists of two IPMC lobes and a proximity sensor, and discuss the design of the control circuitry. We then evaluate the deformation and bending force of the IPMC actuator with various applied signal voltages. We describe a prototype robotic flytrap utilising a proximity sensor to imitate the trigger hairs of the Venus flytrap. We conducted an experiment to assess the feasibility of the biomimetic flytrap. To evaluate grasping ability, we measured the maximum grasping payload with different applied voltages. To enlarge the working area, we integrated biomimetic walking and rotating motion into the robotic Venus flytrap. This paper describes a prototype movable robotic Venus flytrap and evaluates its walking and rotating speeds.     

Development and evaluation of a just-in-time support system

OBJECTIVE: To lay the foundation for a framework of just-in-time support (JITS) for novices dealing with urgent, unfamiliar tasks, and to evaluate a JITS system. BACKGROUND: More than 350,000 people die annually of cardiac arrest in the United States. In response, automated defibrillators are advocated that, unfortunately, do not provide important respiratory support. This paper presents elements of a framework for a JITS system that instructs a lay responder to follow a treatment protocol for integrating respiratory support with the use of an automatic external defibrillator. METHOD: We simulated a medical emergency using a high-fidelity patient simulator and asked participants to care for the patient. Results: When using a paper-based NASA treatment protocol, participants made more errors and took longer to stabilize the injured person than when using the JITS system. CONCLUSION: These findings demonstrate the benefit of a JITS system to instruct novices in unfamiliar tasks. APPLICATION: The JITS system has the potential to improve the treatment outcome of victims of cardiac arrest. The JITS framework can be applied to many situations in which novices deal with urgent tasks without expertise available.     

Real-time cellular-level imaging and medical treatment with a swarm of wireless multifunctional robots

This is a short communication where we present a theoretical model of a swarm of wireless robots that can be used for cellular-level diagnosis and treatment of a variety of life threatening diseases such as cancer. Based on this model, we illustrate a distributed position and orientation tracking algorithm that constructs digitized images from a set of pixels transmitted by the robots of the swarm model that are in motion. Simulation results are also presented.

     

The evaluation of digitally enhanced Indian Remote Sensing Satellite (IRS) data for lithological and structural mapping

The potential of digitally enhanced Indian Remote Sensing Satellite (IRS) data for preparation of lithological and structural maps is explored in two test sites representing different geological settings in parts of Karnataka. Test site 1 is an area around Mattod, which mainly consists of Chitradurga rocks and peninsular gneiss, and the terrain is characterized by bare soil and is devoid of thick vegetation. Test site 2 is an area around the Bababudan belt, which mainly comprised of Bababudan and Chitradurga rocks and peninsular gneiss, and the major part of the area is covered under thick vegetation. Different digital enhancement techniques have been performed, and the expression of different lithological and structural features in the digitally enhanced products were qualitatively assessed. Based on this assessment, only those products which offered additional information than the standard false colour component (FCC) were selected and utilized along with standard FCC in the preparation of geological and structural maps. Attempts have been made to understand the advantages and limitations of using optical remote sensing data for deriving lithological and structural details. The maps created through using digitally enhanced IRS data have helped to revise/modify the existing geological map prepared through conventional method of mapping to considerable extent in terms of refined lithological boundaries, delineation of unmapped rock units, mapping of lineaments, and their networks. The results of the study has clearly brought out the fact that the geological map prepared from the optical remote sensing data at a shorter time interval is broadly comparable to that of the geological map prepared through conventional techniques over several years. Also the study has highlighted that the broad-band spectral resolution optical remote sensing data can be effectively utilized for lithological and structural demarcation, only when the terrain is: (i) exposed with rock outcrops and their weathered derivatives; (ii) characterized by rock units which greatly differ in terms of physico-chemical properties; and (iii) devoid of vegetation cover.     

User interface design for medical imaging workstations: image display and processing

Conventional diagnostic protocols in radiological viewing require the availability of a large display space, as exemplified by the film alternator (or lightbox). However, the recent trend in medical imaging workstation design is the use of relatively small display screens to present radiological images. In the study, the efficacy of an alternatorfilmstrip metaphor for navigating through and displaying the images of a patient study is evaluated. The metaphor relies on the availability of a ‘pictorial directory’ and accommodates the concurrent display of a variable number of images on the screen. In addition the study also evaluated the utility of two commonly available operations for manipulating the appearance of an image: contrast enhancement and image magnification. The study consisted of two experimental variables: availability of a pictorial directory (yes, no) and number of concurrently displayed images (1, 2, or 4). Twelve radiologists viewed two sets of X-ray computed tomography (CT) studies for each of the six treatment conditions and were asked to record an interpretation for each study. Results indicate that interpretations using the two-image display formats took the least amount of time per study, whereas the single-image format with access to the pictorial directory required the longest mean interpretation time. The image magnification capability was generally not found to be useful for image interpretation, whereas the contrast manipulation operations were judged extremely helpful. Implications of the results for medical imaging workstation design are discussed.     

Development and performance evaluation of a new RSS tool for a Web-based system: RSS_PROYECT

The aim of this paper is the development, implementation and subsequent performance evaluation of an innovative tool called RSS_PROYECT, based on the RSS technology for content syndication. At the present, there are several Web content syndicators; although there is no one so far specifically designed for the search of news related to women in risk of social exclusion. They do not even have the characteristics and configuration level of the program syndicator RSS_PROYECT, presented in this paper.To obtain these contents, we used 2 filters (generic and selective), set up by the administrator from the module RSS_PROYECT installed in Joomla!. The generic filter allowed us to do a search of the words included in a series of sources indexed by the user. This filter showed all the sources that contained this word, without exception. In the selective filter, the condition on which the sources would appear would be that all the words of the selective filter were contained in the source.We used different languages such as PHP, MySQL, HTML, XML and Joomla's! Application Program Interface (API). Furthermore, we used the program, Firebug to measure the response speed of the module, both with the selective and generic filters. The results were favorable for the selective filter, and very favorable for the generic filter, concluding that the processing time was short and executed efficiently. Better average times were obtained for RSS_PROYECT with respect to other modules analysed in Joomla!. Nowadays, this tool is used by the Comprehensive Center for Women’s Help (Centro Integral de Ayuda a la Mujer, CIAM) of Valladolid, Spain.     

Compiling SIMULA: a historical study of technological genesis

This article traces the history of the programming language SIMULA from the 1950s into the 1970s, focusing in particular on the formative years between 1962 and 1967. It offers no technical appraisal of the language per se. Rather, it is a sociotechnical analysis aimed at exploring the broader history of the SIMULA project. The article asserts that technological change should be studied in a contextual perspective. Thus the politics surrounding the project and the prehistory of SIMULA are given ample attention