水下便携式三维声纳实时成像系统设计

针对传统三维声纳装置体积庞大、设备沉重,在水下难以灵活作业的问题,设计了水下便携式三维声纳实时成像系统。通过现场可编程门阵列（ FPGA）控制多路信号同步采样,优化波束形成算法大幅提高声纳信号处理速度,同时采用基于低功耗的数字媒体处理器以并行处理方式实时完成三维建模和图像显示。实验结果表明：系统续航时间可达4 h,水下零重量,可在40 m范围内实现角度分辨率1.2°的三维成像,图像刷新率可达25帧/s。     

Digital teaching files in diagnostic imaging

An architecture for digital teaching of radiology targets the next generation of network based education in diagnostic imaging. We developed applications of this system in UCSF hospitals (San Francisco, USA). The networked digital teaching file for medical imaging presented here builds on the availability of existing clinical information infrastructure in hospitals, such as PACS and RIS, and the recent advances in open systems Internet technology. Such a digital teaching file possesses enormous potential to change the practice and landscape of medical imaging education. Countries throughout the world share the goal of improving health care productivity and quality. They must invest not only in clinical services, but also in effective training of a new generation of imaging specialists. A networked digital teaching file provides a cost effective tool for attaining this goal     

计算成像前沿进展

计算成像是融合光学硬件、图像传感器和算法软件于一体的新一代成像技术,突破了传统成像技术信息获取深度(高动态范围、低照度)、广度(光谱、光场、3维)的瓶颈。本文以计算成像的新设计方法、新算法和应用场景为主线,通过综合国内外文献和相关报道来梳理该领域的主要进展。从端到端光学算法联合设计、高动态范围成像、光场成像、光谱成像、无透镜成像、低照度成像、3维成像和计算摄影等研究方向,重点论述计算成像领域的发展现状、前沿动态、热点问题和趋势。端到端光学算法联合设计包括了可微的衍射光学模型、折射光学模型以及基于可微光线追踪的复杂透镜的模型。高动态范围光学成像从原理到光学调制、多次曝光、多传感器融合以及算法等层面阐述不同方法的优点与缺点以及产业应用。光场成像阐述了基于光场的3维重建技术在超分辨、深度估计和3维尺寸测量等方面国内外的研究进展和产业应用,以及光场在粒子测速及3维火焰重构领域的研究进展。光谱成像阐述了当前多通道滤光片,基于深度学习和波长响应曲线求逆问题,以及衍射光栅、多路复用和超表面等优化实现高光谱的获取。无透镜成像包括平面光学元件的设计和优化,以及图像的高质量重建算法。低照度成像包括低照度情...     

Grid enabled magnetic resonance scanners for near real-time medical image processing

This paper presents the initial steps taken to integrate the University of California at San Francisco Radiology Department's magnetic resonance (MR) scanners with its high-performance computing (HPC) grid. The objective is to improve patient care by enabling near real-time, computationally intensive medical image processing, directly at an MR scanner. A graphical software tool is described that was developed to run on the MR scanners for submitting processing jobs to the Departmental grid. The computationally intensive parallel reconstruction and quantification of large, multi-dimensional MR spectroscopic imaging (MRSI) data sets was used as the prototype application for this system. Initial results indicate that real-time processing of medical imaging data on a shared HPC resource is reliable and possible in a clinically acceptable time of less than 5 min. The Department's HPC resource is comprised of hardware owned by multiple research groups at three separate University facilities throughout San Francisco.     

A decision-making algorithm for automatic flow pattern identification in high-speed imaging

A digital image processing algorithm was developed to identify flow patterns in high speed imaging. This numerical tool allows to quantify the fluid dynamic features in compressible flows of relevance in aerospace and space related applications. This technique was demonstrated in a harsh environment with poor image quality and illumination fluctuations. This original pattern recognition tool is based on image binarization and object identification. The geometrical properties of the detected elements are obtained by measuring the characteristics of each object in the binary image. In case of multiple shock waves or shock bifurcations, a “decision-making” algorithm chooses the best shock-wave path, based on the original image intensity and local pattern orientation. The algorithm was successfully used for validation on numerical Schlieren images, where the shock-wave fluctuation was triggered by vortex shedding. The applicability of the algorithm was finally evaluated in two Schlieren imaging studies: at the trailing edge of supersonic airfoils and for hypersonic research. The program correctly identified the fuzzy flow features present in all applications.     

大脑多模态成像技术定量研究进展

现代医学成像技术是脑科学研究和脑疾病诊断的利器,不同模态的成像技术提供不同的信息可协同表征脑部结构和功能。其中定量成像技术着眼于和生理、物理相关的内在参量,旨在提供更精准的信息。本文以正电子发射扫描成像(positron emission tomography,PET)和磁共振成像(magnetic resonance imaging,MRI)两种生物医学成像模态为例,针对性地讨论它们在定量刻画大脑微观结构和功能领域的发展状况,目前尚存的关键技术问题和未来的可能发展方向。围绕定量MRI,从表观参数定量开始,介绍其中的单参数定量的现状和不足,以及目前多参数同时定量的发展动态;围绕微观参数定量,介绍针对髓鞘成像的两大方法,包括多组分T2定量和基于超短回波时间髓鞘直接成像,介绍磁共振定量成像特别是磁共振扩散成像的可比较性和可重复性研究。围绕定量PET,从最广泛的代谢动力学模型——房室模型开始介绍,对生理参数与示踪剂摄取量的关系进行了详细描述,展开到定量的误差来源包括模型选择、图像质量以及输入函数测量误差3个方面进行分析,介绍最新进展包括硬件设备、图像重建方法以及定量分析方法。最后对MRI定量...     

深度学习与生物医学图像分析2020年综述

医学大数据主要包括电子健康档案数据（electronic health record,EHR）、医学影像数据和基因信息数据等,其中医学影像数据占现阶段医学数据的绝大部分。如何将医学大数据应用于临床实践?这是计算机科学研究人员非常关注的问题,医学人工智能提供了一个很好的答案。通过结合医学图像大数据分析方向截至2020年的最新研究进展,以及医学图像大数据分析领域最近的工作,梳理了当前在医学图像领域以核磁共振影像、超声影像、病理和电信号为代表的4个子领域以及部分其他方向使用深度学习进行图像分析的方法理论和主要流程,对不同算法进行结果评价。本文分析了现有算法的优缺点以及医学影像领域的重难点,介绍了智能成像和深度学习在大数据分析以及疾病早期诊断领域的应用,同时展望了本领域未来的发展热点。深度学习在医学影像领域发展迅速,发展前景广阔,对疾病的早期诊断有重要作用,能有效提高医生工作效率并减轻负担,具有重要的理论研究和实际应用价值。     

Supervised restoration of degraded medical images using multiple-point geostatistics

Reducing noise in medical images has been an important issue of research and development for medical diagnosis, patient treatment, and validation of biomedical hypotheses. Noise inherently exists in medical and biological images due to the acquisition and transmission in any imaging devices. Being different from image enhancement, the purpose of image restoration is the process of removing noise from a degraded image in order to recover as much as possible its original version. This paper presents a statistically supervised approach for medical image restoration using the concept of multiple-point geostatistics. Experimental results have shown the effectiveness of the proposed technique which has potential as a new methodology for medical and biological image processing.     

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。     

基于Morphology和SVM对细胞组织的分析

随着计算机技术、成像技术、图像处理技术、医学形态学研究的发展,计算机辅助图像处理成为了医学研究的得力工具。论文利用数学形态学(Morphology)和小波变换(Wavelet)对细胞图像进行分析和识别,充分发挥形态学善于处理图像形态的优势和小波变换系数反映细节的特点,通过图像预处理、分割和特征提取,最后通过支持向量机分类器(SVM)来完成细胞状态的自动分类和识别,试验表明该方法具有较高的准确率和可靠性。     

深度学习在口腔医学影像中的应用与挑战

口腔医学影像是进行临床口腔疾病检测、筛查、诊断和治疗评估的重要工具,对口腔影像进行准确分析对于后续治疗计划的制定至关重要。常规的口腔医学影像分析依赖于医师的水平和经验,存在阅片效率低、可重复性低以及定量分析欠缺的问题。深度学习可以从大样本数据中自动学习并获取优良的特征表达,提升各类机器学习任务的效率和性能,目前已广泛应用于医学影像分析处理的各类任务之中。基于深度学习的口腔医学影像处理是目前的研究热点,但由于口腔医学领域内在的特殊性和复杂性,以及口腔医学影像数据样本量通常较小的问题,给深度学习方法在相关学习任务和场景的应用带来了新的挑战。本文从口腔医学影像领域常用的二维X射线影像、三维点云/网格影像和锥形束计算机断层扫描影像3种影像出发,介绍深度学习技术在口腔医学影像处理及分析领域应用的思路和现状,分析了各算法的优缺点及该领域所面临的问题和挑战,并对未来的研究方向和可能开展的临床应用进行展望,以助力智慧口腔建设。     

A hospital integrated framework for multimodality image basemanagement

The trend in healthcare information technology is increasingly digital and multimedia oriented. The next generation of health care information systems will consist of a vast network of heterogeneous, autonomous, and distributed imaging scanners, databases, information systems, knowledge intensive applications, and large quantities of multimedia medical data. A key challenge facing system researchers and builders is to provide a new organizational framework that can integrate this varied collection of resources into what appears to be a uniform and logical conglomeration of data and knowledge store in order to increase the availability of global or previously nonaccessible information and to address demanding new information processing requirements for diverse image-assisted medical applications. The purpose of this paper is to present the authors' research toward the development of a hospital integrated framework of multimodality image base management (MIBM) for digital radiology of the future. This evolutionary framework consists of three hierarchical components: a hospital-integrated picture archiving and communication system (HI-PACS), a medical image database system (MIDS), and a set of image-based medical applications that relies on the support of MIDS and PACS. In this paper, the authors describe the system architecture, guiding principles, and design specifications of HI-PACS and MIDS and illustrate their functions and capabilities with three implemented applications, namely, patient folder workflow, distributed object management, and multimodality imaging studies. In addition, the authors conclude their findings with a summary of challenges and research directions     

Applications of 'TissueQuant'- a color intensity quantification tool for medical research

This paper demonstrates the use of TissueQuant - an image analysis tool for quantification of color intensities which was developed for use in medical research where the stained biological specimen such as tissue or antigen needs to be quantified. TissueQuant provides facilities for user interaction to choose and quantify the color of interest and its shades. Gaussian weighting functions are used to provide a color score which quantifies how close the shade is to the user specified reference color. We describe two studies in medical research which use TissueQuant for quantification. The first study evaluated the effect of petroleum-ether extract of Cissus quadrangularis (CQ) on osteoporotic rats. It was found that the analysis results correlated well with the manual evaluation, p < 0.001. The second study evaluated the nerve morphometry and it was found that the adipose and non adipose tissue content was maximum in radial nerve among the five nerves studied.     

Applications of deformable models for in-dopth analysis and feature extraction from medical images—A review

Medical imaging plays important role for the practice of medicine globally and is rapidly increasing and getting sophisticated day by day. But accurate, fully automatic medical image analysis continues to be an elusive ideal for quantitative exploitation for diagnosis and therapy. The spatial relations between these anatomical structures as well as the biological shape variations are observed over a representative population of individuals. Deformable models (DMs) are being used in digital image analysis to maintain essential characteristics of image shape and intensity while accommodating fluctuations. Among model-based techniques, DMs offer a unique and powerful approach to image analysis that combines geometry, physics, and approximation theory. DMs are highly insightful interactive methods that allow medical scientists and practitioners to bring their expertise to bear on the model-based image interpretation task whenever necessary. The paper reviews the application of deformable models as a capable and robustly applied digital medical image analysis technique of the human body.     

Interactive histology of large-scale biomedical image stacks

Histology is the study of the structure of biological tissue using microscopy techniques. As digital imaging technology advances, high resolution microscopy of large tissue volumes is becoming feasible; however, new interactive tools are needed to explore and analyze the enormous datasets. In this paper we present a visualization framework that specifically targets interactive examination of arbitrarily large image stacks. Our framework is built upon two core techniques: display-aware processing and GPU-accelerated texture compression. With display-aware processing, only the currently visible image tiles are fetched and aligned on-the-fly, reducing memory bandwidth and minimizing the need for time-consuming global pre-processing. Our novel texture compression scheme for GPUs is tailored for quick browsing of image stacks. We evaluate the usability of our viewer for two histology applications: digital pathology and visualization of neural structure at nanoscale-resolution in serial electron micrographs.     

Neuromolecularware and its application to pattern recognition

Unlike computer systems, organisms have high adaptability in dealing with environmental changes or noise. The ability to evolve, self-organizing dynamics, and a closed structure–function relationship are the three principle features embedded in biological structures that provide great malleability to environmental change. Computer systems have fast processing speed for performing heavy computational tasks. One of the objectives in this research is to capture these three biological features and implement them onto a digital circuit. The proposed hardware (called neuromolecular hardware), is the integration of inter- and intraneuronal information processing applied to the pattern recognition problem domain. This approach was tested on the Quartus II system, a simulation tool for digital circuits. The experimental result showed good self-organizing capability in selecting significant bits for differentiating patterns and insignificant bits for tolerating noise. The proposed digital circuit also exhibited a closed structure–function relationship. This implied that this hardware embraced an adaptive fitness landscape that facilitated processing spatiotemporal information.     

Non-destructive on-chip imaging flow cell-sorting system for on-chip cellomics

We have developed a non-destructive imaging flow cell-sorting system using an ultra-high-speed camera (shutter speed of 1/10,000 s) with a real-time image analysis unit and a poly(methyl methacrylate) (PMMA)-based disposable microfluidic chip for single-cell-based on-chip cellomics. It has a 3-D micropipetting device that supports fully automated sorting and collection of samples. The entire fluidic system is implemented in a disposable plastic chip, enabling biological samples to be lined up in a laminar flow using hydrodynamic focusing. Its optical system enables direct observation-based cell identification using specific image indexes and phase-contrast/fluorescence microscopy, real-time image processing. It has a non-destructive, wider dynamic range, sorting procedure using mild electrostatic force in a laminar flow; agarose gel electrodes are used to prevent electrode loss and electrolysis bubble formation. The microreservoir used for recultivating collected target cells is contamination-free. An integrated ultra-high-speed droplet polymerase chain reaction measurement module is used for DNA/mRNA analysis of the collected target cells. This system was used to separate cardiomyocyte cells from a mixture of various cells. All the operations were automated using the 3-D micropipetting device. The results demonstrate that this imaging flow cell-sorting system is practically applicable for biological research and clinical diagnosis.     

机器学习在术中光学成像技术中的应用研究

术中光学成像技术的兴起为临床手术提供了更加便捷和直观的观察手段。传统的术中光学成像方法包括开放式光学成像和术中腔镜、内镜成像等,这些方法保障了临床手术的顺利进行,同时也促进了微创手术的发展。随后发展起来的术中光学成像技术还有窄带腔镜成像、术中激光共聚焦显微成像和近红外激发荧光成像等。术中光学成像技术可以辅助医生精准定位肿瘤、快速区分良恶性组织和检测微小病灶等,在诸多临床应用领域表现出了较好的应用效果。但术中光学成像技术也存在成像质量受限、缺乏有力的成像分析工具,以及只能成像表浅组织的问题。机器学习的加入,有望突破瓶颈,进一步推动术中光学成像技术的发展。本文针对术中光学成像技术,对机器学习在这一领域的应用研究展开调研,具体包括：机器学习对术中光学成像质量的优化、辅助术中光学成像的智能分析,以及辅助基于术中光学影像的3维建模等内容。本文对机器学习在术中光学成像领域的应用进行总结和分析,特别叙述了深度学习方法在该领域的应用前景,为后续的研究提供更宽泛的思路。     

Scalable edge enhancement with automatic optimization for digital radiographic images

The ability to resolve fine picture detail is of paramount importance in medical imaging systems for viewing small tissue, bone structure and anatomy in X-ray images. In this paper, we present a new digital radiographic image processing system with the property of scalability and adaptability. (i) A new automatic optimization algorithm is proposed for display. (ii) An adaptive detection of a region-of-interest is developed. (iii) A “scalable edge enhancement algorithm” is proposed to improve the image quality for showing subtle structures in digital radiographic images. The advantage of the proposed method is demonstrated through experiments on 200 digital X-ray images and 50 CT images, in which different parts of human body structures are captured.     

Overview Paper: Intelligent color imaging*

Abstract— With the 600th anniversary of Johann Gutenberg's birth in 2000 A.D., we should look back at the historical significance of letterpress technology and take a step forward into the new age of color imaging. Currently, digital‐imaging technology plays a leading role in visual communications, but meets severe challenges to satisfy human vision. It is essential to consider what human vision “sees” in order to capture, store, transmit, and reproduce a truly realistic image just as human vision. Advances not only in high‐precision and high‐definition digital media but also in intelligent image‐processing technologies will be helpful in producing more aesthetic and pleasant imaging. This paper briefly introduces “intelligent” processing towards that goal by using region‐based spatially variant, and scene‐referred approaches.