Tom Weidong Cai

Content-Based Medical Image Retrieval

Publisher Summary This chapter introduced content-based image retrieval (CBIR) and its key components. CBIR methods support full retrieval by visual content/properties of images, by retrieving image data at a perceptual level with objective and quantitative measurements of the visual content and integration of image processing, pattern recognition, and computer vision. The four key issues in any CBIR system are feature extraction, similarity comparison, the indexing scheme, and the interactive query interface. Feature extraction is the basic and most important component of the CBIR system. The potential benefits of content-based medical image retrieval (CBMIR) range from clinical decision support to medical education and research. Diagnosis by comparing past and current medical images associated with pathological conditions has become one of the primary approaches in case-based reasoning. The CBMIR techniques are designed mainly for anatomical images that capture human anatomy at different levels and provide primarily structural information. The ultimate goal for CBMIR is to find medically meaningful similar cases. Sometimes similar visual features in images may not imply similar diagnoses or symptoms, and vice versa. Combining visual features in the image with domain knowledge to reach the right subset of relevant cases in the database is a key to the success of CBMIR. It would therefore be useful as a training tool for medical students, residents, and researchers to browse and search large collections of disease-related illustrations using their visual attributes.

New Block-Based Motion Estimation for Sequences with Brightness Variation and Its Application to Static Sprite Generation for Video Compression

In this brief, a new local motion estimator is proposed which can accurately estimate motion activities under varying strong brightness conditions. The proposed estimator makes use of a new block division technique which manages practically to get rid of the adverse influence caused by brightness changes between frames. We also propose a new static sprite coding system using the proposed local motion estimator. The system is characterized not only with the features of accurate motion estimation under varying brightness conditions, but also possesses the capability of coding the brightness variability of the background scene using a single layered sprite image. Experimental results show that the resulting static sprite coding system improves the PSNR by 6.32 dB as compared with the conventional static sprite coding system when the background scenes of the video sequences involve strong brightness variations in the spatial and time domains.

Automatic 3D temporal kinetics segmentation of dynamic emission tomography image using adaptive region growing cluster analysis

The primary goal of medical image segmentation is to partition the raw image into region of interests (ROIs) matching the anatomical localization of objects of interest in 2D or 3D space. The traditional method of ROI delineation (or segmentation) for the analysis of dynamic emission tomography is the manual placement of ROIs by the operator. However this approach is operator dependent, time-consuming and may lack good reproducibility. Quantitative positron emission tomography (PET) studies can provide measurements of dynamic physiological and biochemical processes in humans through the use of temporal kinetics available. However, due to the relatively poor spatial resolution and high noise levels, partitioning of ROIs is limited. In this paper, the use of a novel knowledge-based approach to segmentation of clinical PET studies using automatic seed selection for adaptive region growing based on Euclidean distance between the local tissue time-activity curves (TTAC) of the voxels is proposed.

Multimedia for Future Health—Smart Medical Home

Publisher Summary Multimedia technologies are enabling more comprehensive and intuitive uptake of information in a wide range of fields that have a direct impact on our life, particularly in entertainment, education, work, and health. The core components behind these multimedia technologies are human-centered multimedia services, which combine many fields of information technology including computing, telecommunication, databases, mobile devices, sensors, and virtual/augmented reality systems. Human-centered multimedia services are built upon three key research pillars, namely, human-computer interaction (HCI), multimedia delivery, and multimedia data management. This chapter presents the latest research and development in multimedia technologies and the transition of these technologies into health care products for the smart medical home. It is subdivided into two parts: enabling multimedia technologies, and applications involving multimedia technologies in biomedicine. The aim of HCI is to mimic human-human interactions. The primary aim of medical homes is to develop an integrated health system that is personalized to an individual’s home. This technology will allow consumers, in the privacy and comfort of their own homes, to maintain health, detect the onset of disease, and manage symptoms. The smart medical home has the potential to delay or partially remove the dependence on retirement nursing homes and thereby extend the person’s quality of life. Incorporating smart medical devices into homes can potentially make a strong and positive impact on the lives of persons with physical disabilities and those with chronic diseases.

Data Visualization and Display

Publisher Summary This chapter introduces the visualization and display technologies that are currently being employed in clinical applications, as well as techniques that have the potential to enhance and provide improved diagnostic capabilities. In 3D visualization of biomedical data, there are four common techniques that may be used to maximize the visual information, and these are surface rendering, direct volume rendering, texture-based volume rendering, and multivariate rendering. In 3D, navigation is based on movement of the volume with an additional z-axis. Data with even greater dimensions, such as multivariate data, demand additional interactive features and input devices for efficient control. Enhancement and manipulation techniques applied to 2D slice images, such as filtering, segmentation, and geometrical transformation, can also be extended and applied to 3D image volumes. When extended into the 3D domain, additional information is available, and this therefore generally allows for more accurate methods. It also highlights some of the solutions for optimizing volume-rendering visualization of large biomedical data. In the visualization of dual-modality PET/CT, the PET images provide high sensitivity in tumor detection and tissue characterization. The effectiveness of all diagnostic imaging visualization is affected by the quality of the display devices. Modern visualization and display technologies enable the development of clinical applications that provide new approaches for interacting with and interpreting biomedical data, such as in virtual biopsy, motion activity visualization, and applications in radiotherapy planning.

Quarantine MAP reconstruction of PET/CT data using dual priors

Maximum a posteriori (MAP) reconstruction makes use of an anatomical prior from CT or MRI imaging to enforce smoothness of reconstructed PET images while preserving anatomical edges. The tendency of this technique to smooth parts of the image between anatomical boundaries may reduce the detectability of functional lesions if, as is commonly the case, the edges of these lesions do not conform to anatomical boundaries. We have investigated the use of a functional prior in addition to an anatomical prior to improve the detection and quantification of lesions in PET imaging. We introduce a new parameter, Q, which controls the weight, β, of the functional prior on a spatially-variant basis, to enable a reduction of the smoothing effect in regions containing lesions. Such regions constitute the functional prior. They can be defined, for example, by applying a threshold to a preliminary reconstructed PET image. They are quarantined from the smoothing of the standard MAP algorithm, and subjected to a lesser degree of smoothing as determined by the combined effects of Q and β. We call this dual-prior technique quarantine MAP reconstruction (QMAP). Thus the method alters the degree of smoothing in specific parts of the image with the aim of enhancing lesion detectability. We have compared the QMAP algorithm in computer simulations with standard One-Step-Late (OSL) MAP reconstruction and OSL-MAP with CT prior information. QMAP provided better lesion contrast than the other algorithms, without altering the properties of other parts of the image.

Content access and distribution of multimedia medical data in E-health

E-health is greatly impacting on information distribution and availability within the health services, hospitals and to the public. Previous research has addressed the development of system architectures with the aim of integrating the distributed and heterogeneous medical information systems. Easing the difficulties in the sharing and management of multimedia medical data and the timely accessibility to these data are critical needs for health care providers. We have proposed a client-server agent that integrates and allows a portal to every permitted information system of the hospital that consists of picture archiving and communication systems (PACS), radiology information system (RIS) and hospital information system (HIS) via the intranet and the Internet. Our proposed agent enables remote access into the usually closed information system of the hospital and a server that manages all the multimedia medical data and allows for in-depth and complex search queries for content access and automatic creation of patient reports for distribution.

Semi-supervised manifold learning with affinity regularization for Alzheimer's disease identification using positron emission tomography imaging.

Dementia, Alzheimers disease (AD) in particular is a global problem and big threat to the aging population. An image based computer-aided dementia diagnosis method is needed to providing doctors help during medical image examination. Many machine learning based dementia classification methods using medical imaging have been proposed and most of them achieve accurate results. However, most of these methods make use of supervised learning requiring fully labeled image dataset, which usually is not practical in real clinical environment. Using large amount of unlabeled images can improve the dementia classification performance. In this study we propose a new semi-supervised dementia classification method based on random manifold learning with affinity regularization. Three groups of spatial features are extracted from positron emission tomography (PET) images to construct an unsupervised random forest which is then used to regularize the manifold learning objective function. The proposed method, stat-of-the-art Laplacian support vector machine (LapSVM) and supervised SVM are applied to classify AD and normal controls (NC). The experiment results show that learning with unlabeled images indeed improves the classification performance. And our method outperforms LapSVM on the same dataset.

Image data compression and storage

Picture compression is an important tool in the modern digital world. Over the years, the move toward the digital media has led to the proliferation of digital picture compression systems. This is particularly noticeable in the entertainment industry, consumer electronics, and security/surveillance systems. However, data compression is becoming increasingly important in biomedical imaging applications as well, due to the increased popularity of digital biomedical imaging systems, the constant improvement of image resolution, and the practical need for online sharing of information through networks. Picture compression came about following the advent of analog television broadcasting. Initial methods of limiting signal bandwidth for transmission were relatively simple. They included subsampling to lower picture resolution and/or interlacing of television pictures into alternate Welds in alternate picture frames. With the introduction of color television, subsampling was extended to the color channels as well [1]. The digital picture is seen as a natural migration from the analog picture. Hence, it is felt that digital picture compression is, in many ways, a natural evolution of analog compression. While this is true in some sense, the nature of digital signals and of analog signals is quite different. Consequently, the methods for compressing digital and analog pictures are distinct from one another. This chapter will first provide a basic introduction to digital picture compression, focusing on general concepts and methods, then introduce some advanced data compression techniques. These techniques are used in noisy medical image data sets with high compression ratios (CRs) and improved image quality, which have pioneered in biomedical diagnostically lossless data compression research. An extended discussion of classical data compression can be found in [2-8], while the new research in diagnostically lossless data compression can be found in the references given in the later sections.

Content-Based Retrieval for Medical Data

The recent information explosion has led to massively increased demand for multimedia data storage in integrated database systems. Content-based retrieval is an important alternative and complement to traditional keyword-based searching for multimedia data, and can greatly enhance information management. However, current content-based image retrieval techniques have some deficiencies when applied in the medical imaging domain. Many of the proposed techniques for content-based retrieval of medical data use features or patterns specific to medical images. In this chapter, we address content-based retrieval techniques for the following types of medical data: one-dimensional ECG signals (Section 17.2); two-dimensional X-ray projection images (Section 17.3); three-dimensional CT/MRI volume images (Section 17.4); and four-dimensional PET / SPECT dynamic images (Section 17.5). Finally, a summary is given in Section 17.6.