We have witnessed great interest and a wealth of promise in content-based image retrieval as an emerging technology. While the last decade laid foundation to such promise, it also paved the way for a large number of new techniques and systems, got many new people involved, and triggered stronger association of weakly related fields. In this article, we survey almost 300 key theoretical and empirical contributions in the current decade related to image retrieval and automatic image annotation, and in the process discuss the spawning of related subfields. We also discuss significant challenges involved in the adaptation of existing image retrieval techniques to build systems that can be useful in the real world. In retrospect of what has been achieved so far, we also conjecture what the future may hold for image retrieval research.

Abstract Recurrent spontaneous abortion (RSA) is de-fined as three or more consecutive pregnancy losses prior to the 20th week of gestation. The etiology of recurrent spontaneous abortion is often unclear and may be multifactorial, with much controversy regarding diag-nosis and treatment. Reasonably accepted etiologic causes include, genetics, anatomical, endocrine, placen-tal anomalies, hormonal problems, infection, smoking and alcohol consumption, exposure to environmental factors, psychological trauma and stressful life event, certain coagulation and immunoregulatory protein de-fects. Detection of an abnormality in any of these areas may result into specific therapeutic measures, with varying degrees of success. However, the majority of cases of RSA remains unexplained and is found to be associated with certain autoimmune (APA, ANA, ACA, ATA, AECA) and alloimmune (APCA, Ab2, MLR-Bf) antibodies that may play major role in the immunologic failure of pregnancy and may lead to abortion. Alter-ation in the expression of HLA-G molecules, T-helper-1 (Th-1) pattern of cytokines and natural killer (NK) cells activity may also induce abortion. Various forms of treatment like antithrombotic therapies such as aspirin and heparin, intravenous immunoglobulin (IVIg) ther-apy, immunotherapy with paternal lymphocytes and vitamin D3 therapy are effective mode of treatment for unexplained cause of fetal loss in women with RSA.

A probabilistic framework is presented that enables image registration, tissue classification, and bias correction to be combined within the same generative model. A derivation of a log-likelihood objective function for the unified model is provided. The model is based on a mixture of Gaussians and is extended to incorporate a smooth intensity variation and nonlinear registration with tissue probability maps. A strategy for optimising the model parameters is described, along with the requisite partial derivatives of the objective function.

The techniques available for the interrogation and analysis of neuroimaging data have a large influence in determining the flexibility, sensitivity and scope of neuroimaging experiments. The development of such methodologies has allowed investigators to address scientific questions which could not previously be answered and, as such, has become an important research area in its own right. In this paper, we present a review of the research carried out by the Analysis Group at the Oxford Centre for Functional MRI of the Brain (FMRIB). This research has focussed on the development of new methodologies for the analysis of both structural and functional magnetic resonance imaging data. The majority of the research laid out in this paper has been implemented as freely available software tools within FMRIB’s Software Library (FSL). 1

Characterizing the performance of image segmentation approaches has been a persistent challenge. Performance analysis is important since segmentation algorithms often have limited accuracy and precision. Interactive drawing of the desired segmentation by human raters has often been the only acceptable approach, and yet suffers from intrarater and inter-rater variability. Automated algorithms have been sought in order to remove the variability introduced by raters, but such algorithms must be assessed to ensure they are suitable for the task. The performance of raters...

Abstract -We present a technique for automatically assigning a neuroanatomical label to each location on a cortical surface model based on probabilistic information estimated from Introduction Techniques for labeling geometric features of the cerebral cortex are useful for analyzing a variety of functional and structural neuroimaging data Unfortunately, despite their potential utility, cortical parcellations are not commonly used in the neuroimaging community due to the difficult and time-consuming nature of the task of manually parcellating the entire cortex from high-resolution MRI images. While many techniques exist for labeling parts of the cortex The inclusion of prior information is a critical feature of a cortical parcellation algorithm. The reason this is the case is that the divisions that are useful in a cortical parcellation scheme can be based on properties of the brain other than cortical geometry. Many disparate pieces of information, such as knowledge of structure-function relationship, and cytoarchitectonic or receptor labeling properties of regions, may be used by a neuroanatomist in generating a cortical parcellation -information that is not directly available to an automated parcellation procedure from magnetic resonance imaging (MRI) data. A trivial example of this is the fact that many unbroken sulci change names as lobar boundaries are crossed. In addition, functional heterogeneity can make it desirable for adjacent areas to be assigned different labels despite the absence of macroscopic cortical features to distinguish them. Conversely, there are situations in which the secondary folding structure of the cortex (i.e. folds within folds) carries information, such as the location of the hand area in primary motor cortex, which is well predicted by a posterior-pointing secondary fold in the precentral gyrus Various approaches have been taken to the problem of labeling of cortical features. For example, Sandor and Leahy (Sandor and Leahy, 1997) use a manually labeled atlas brain, which is then warped into correspondence with an individual subject's surface model. The individual subject surface then inherits the labels from the atlas. Similar surface-fitting approaches have been developed by Here we present a technique for using manually labeled data as the basis for an automated parcellation procedure, using an intrinsically cortical coordinate system to store prior statistics and classconditional densities . The use of the manual parcellation as a training set allows the technique to incorporate neuroanatomical convention into the parcellation in regions in which geometry alone is not predictive of a parcellation label. The procedure models the parcellation labels as a first order anisotropic nonstationary Markov random field (MRF), allowing it to capture the spatial relationships between parcellation units that are present in the training set. The anisotropy separates label probabilities in the first and second principal curvature directions, in order to encode the increased probability of differing labels in the direction of maximal curvature. This type of modeling allows the procedure to automatically encode information such as "precentral gyrus frequently neighbors central sulcus in the direction of high curvature (moving across the sulcus), but not in the direction of low curvature (along the banks of the sulcus)". This type of parcellation has commonly been generated by a having a trained anatomist or technician manually label some or all of the structures in the cortex, a procedure that can take up to a week for high-resolution images. Here, we use the results of the manual labeling using the validated techniques of the Center for Morphometric Analysis (CMA)

The anatomical connectivity pattern of a brain region determines its function 1 . Although invasive tracer studies have produced a large body of evidence concerning connectivity patterns in non-human animals 2-4 , direct information concerning brain connections in humans is very limited. Injection of fluorescent dyes post mortem allows tracing of tracts, but only for distances of tens of millimeters 5 . Longer-distance connections can be investigated by dissection of major tracts or histological studies of remote degeneration following a focal lesion 6 , but such work is based on a relatively small number of informative patients. A specific, important focus for investigation is the thalamus because nearly all incoming information to the cortex is routed through this deep gray-matter structure. The thalamus is subdivided into cytoarchitectonically distinct nuclei which have different patterns of anatomical connectivity that are well characterized for non-human animals Diffusion imaging characterizes the apparent diffusion properties of water Here, using a probabilistic tractography algorithm, we were able to infer anatomical connectivity that progresses fully into gray matter. We thus provide a comprehensive description of the connections between thalamus and cortex in the human brain in vivo. An additional result of this approach is the discrimination of human thalamic subregions on the basis of their connections with the cortex. RESULTS We used a fully automated probabilistic tractography algorithm (see Methods) to form connectivity distributions from individual voxels within the thalamus of a single subject. From these distributions, we traced pathways all the way to the cortex Commonly connected thalamic subregions We segmented the cortex into large, anatomically defined regions (see Methods) corresponding to known connection areas of the major thalamic nuclear groups in non-human primates

The last decade has witnessed great interest in research on content-based image retrieval. This has paved the way for a large number of new techniques and systems, and a growing interest in associated fields to support such systems. Likewise, digital imagery has expanded its horizon in many directions, resulting in an explosion in the volume of image data required to be organized. In this paper, we discuss some of the key contributions in the current decade related to image retrieval and automated image annotation, spanning 120 references. We also discuss some of the key challenges involved in the adaptation of existing image retrieval techniques to build useful systems that can handle real-world data. We conclude with a study on the trends in volume and impact of publications in the field with respect to venues/journals and sub-topics.

Background. The present study examined whether aerobic fitness training of older humans can increase brain volume in regions associated with age-related decline in both brain structure and cognition. Methods. Fifty-nine healthy but sedentary community-dwelling volunteers, aged 60–79 years, participated in the 6-month randomized clinical trial. Half of the older adults served in the aerobic training group, the other half of the older adults participated in the toning and stretching control group. Twenty young adults served as controls for the magnetic resonance imaging (MRI), and did not participate in the exercise intervention. High spatial resolution estimates of gray and white matter volume, derived from 3D spoiled gradient recalled acquisition MRI images, were collected before and after the 6-month fitness intervention. Estimates of maximal oxygen uptake (VO2) were also obtained. Results. Significant increases in brain volume, in both gray and white matter regions, were found as a function of fitness training for the older adults who participated in the aerobic fitness training but not for the older adults who participated in the stretching and toning (nonaerobic) control group. As predicted, no significant changes in either gray or white matter volume were detected for our younger participants. Conclusions. These results suggest that cardiovascular fitness is associated with the sparing of brain tissue in aging humans. Furthermore, these results suggest a strong biological basis for the role of aerobic fitness in maintaining and enhancing central nervous system health and cognitive functioning in older adults.

& The Open Access Series of Imaging Studies is a series of mag-netic resonance imaging data sets that is publicly available for study and analysis. The initial data set consists of a cross-sectional collection of 416 subjects aged 18 to 96 years. One hundred of the included subjects older than 60 years have been clinically diagnosed with very mild to moderate Alzheimer’s disease. The subjects are all right-handed and include both men and women. For each subject, three or four individual T1-weighted magnetic resonance imaging scans obtained in single imaging sessions are included. Multiple within-session acquisitions provide extremely high contrast-to-noise ratio, making the data amenable to a wide range of analytic approaches including automated computational analysis. Additionally, a reliability data set is included containing 20 subjects without dementia imaged on a subsequent visit within 90 days of their initial session. Automated calculation of whole-brain volume and estimated total intracranial volume are presented to demonstrate use of the data for measuring differ-ences associated with normal aging and Alzheimer’s disease. &