On the convergence of EM-like algorithms for image segmentation using Markov random fields.

Inference of Markov random field images segmentation models is usually performed using iterative methods which adapt the well-known expectation-maximization (EM) algorithm for independent mixture models. However, some of these adaptations are ad hoc and may turn out numerically unstable. In this paper, we review three EM-like variants for Markov random field segmentation and compare their convergence properties both at the theoretical and practical levels. We specifically advocate a numerical scheme involving asynchronous voxel updating, for which general convergence results can be established. Our experiments on brain tissue classification in magnetic resonance images provide evidence that this algorithm may achieve significantly faster convergence than its competitors while yielding at least as good segmentation results.

Mixed spectral-structural classification of very high resolution images with summation kernels

In this paper, mixed spectral-structural kernel machines are proposed for the classification of very-high resolution images. The simultaneous use of multispectral and structural features (computed using morphological filters) allows a significant increase in classification accuracy of remote sensing images. Subsequently, weighted summation kernel support vector machines are proposed and applied in order to take into account the multiscale nature of the scene considered. Such classifiers use the Mercer property of kernel matrices to compute a new kernel matrix accounting simultaneously for two scale parameters. Tests on a Zurich QuickBird image show the relevance of the proposed method : using the mixed spectral-structural features, the classification accuracy increases of about 5%, achieving a Kappa index of 0.97. The multikernel approach proposed provide an overall accuracy of 98.90% with related Kappa index of 0.985.

Image guided critical screw placement precision in the cervical spine : important or incidental? : results of a series of 91 consecutive cases using OVEM

Study design: A retrospective study of image guided cervical implant placement precision. Objective: To describe a simple and precise classification of cervical critical screw placement. Summary of Background Data: "Critical" screw placement is defined as implant insertion into a bone corridor which is surrounded circumferentially by neurovascular structures. While the use of image guidance has improved accuracy, there is currently no classification which provides sufficient precision to assess the navigation success of critical cervical screw placement. Methods: Based on postoperative clinical evaluation and CT imaging, the orthogonal view evaluation method (OVEM) is used to classify screw accuracy into grade I (no cortical breach), grade la (screw thread cortical breach), grade II (internal diameter cortical breach) and grade III (major cortical breach causing neural or vascular injury). Grades II and III are considered to be navigation failures, after accounting for bone corridor / screw mismatch (minimal diameter of targeted bone corridor being smaller than an outer screw diameter). Results: A total of 276 screws from 91 patients were classified into grade I (64.9%), grade la (18.1%), and grade II (17.0%). No grade III screw was observed. The overall rate of navigation failure was 13%. Multiple logistic regression indicated that navigational failure was significantly associated with the level of instrumentation and the navigation system used. Navigational failure was rare (1.6%) when the margin around the screw in the bone corridor was larger than 1.5 mm. Conclusions: OVEM evaluation appears to be a useful tool to assess the precision of critical screw placement in the cervical spine. The OVEM validity and reliability need to be addressed. Further correlation with clinical outcomes will be addressed in future studies.

Effect of training-sample size and classification difficulty on the accuracy of genomic predictors.

Introduction: As part of the MicroArray Quality Control (MAQC)-II project, this analysis examines how the choice of univariate feature-selection methods and classification algorithms may influence the performance of genomic predictors under varying degrees of prediction difficulty represented by three clinically relevant endpoints. Methods: We used gene-expression data from 230 breast cancers (grouped into training and independent validation sets), and we examined 40 predictors (five univariate feature-selection methods combined with eight different classifiers) for each of the three endpoints. Their classification performance was estimated on the training set by using two different resampling methods and compared with the accuracy observed in the independent validation set. Results: A ranking of the three classification problems was obtained, and the performance of 120 models was estimated and assessed on an independent validation set. The bootstrapping estimates were closer to the validation performance than were the cross-validation estimates. The required sample size for each endpoint was estimated, and both gene-level and pathway-level analyses were performed on the obtained models. Conclusions: We showed that genomic predictor accuracy is determined largely by an interplay between sample size and classification difficulty. Variations on univariate feature-selection methods and choice of classification algorithm have only a modest impact on predictor performance, and several statistically equally good predictors can be developed for any given classification problem.

Anti-tumor necrosis factor drug survival in axial spondyloarthritis is independent of the classification criteria.

To compare the impact of meeting specific classification criteria [modified New York (mNY), European Spondyloarthropathy Study Group (ESSG), and Assessment of SpondyloArthritis international Society (ASAS) criteria] on anti-tumor necrosis factor (anti-TNF) drug retention, and to determine predictive factors of better drug survival. All patients fulfilling the ESSG criteria for axial spondyloarthritis (SpA) with available data on the axial ASAS and mNY criteria, and who had received at least one anti-TNF treatment were retrospectively retrieved in a single academic institution in Switzerland. Drug retention was computed using survival analysis (Kaplan-Meier), adjusted for potential confounders. Of the 137 patients classified as having axial SpA using the ESSG criteria, 112 also met the ASAS axial SpA criteria, and 77 fulfilled the mNY criteria. Drug retention rates at 12 and 24 months for the first biologic therapy were not significantly different between the diagnostic groups. Only the small ASAS non-classified axial SpA group (25 patients) showed a nonsignificant trend toward shorter drug survival. Elevated CRP level, but not the presence of bone marrow edema on magnetic resonance imaging (MRI) scans, was associated with significantly better drug retention (OR 7.9, ICR 4-14). In this cohort, anti-TNF drug survival was independent of the classification criteria. Elevated CRP level, but not positive MRI, was associated with better drug retention.

Generating controlled image sets in cognitive neuroscience research.

The investigation of perceptual and cognitive functions with non-invasive brain imaging methods critically depends on the careful selection of stimuli for use in experiments. For example, it must be verified that any observed effects follow from the parameter of interest (e.g. semantic category) rather than other low-level physical features (e.g. luminance, or spectral properties). Otherwise, interpretation of results is confounded. Often, researchers circumvent this issue by including additional control conditions or tasks, both of which are flawed and also prolong experiments. Here, we present some new approaches for controlling classes of stimuli intended for use in cognitive neuroscience, however these methods can be readily extrapolated to other applications and stimulus modalities. Our approach is comprised of two levels. The first level aims at equalizing individual stimuli in terms of their mean luminance. Each data point in the stimulus is adjusted to a standardized value based on a standard value across the stimulus battery. The second level analyzes two populations of stimuli along their spectral properties (i.e. spatial frequency) using a dissimilarity metric that equals the root mean square of the distance between two populations of objects as a function of spatial frequency along x- and y-dimensions of the image. Randomized permutations are used to obtain a minimal value between the populations to minimize, in a completely data-driven manner, the spectral differences between image sets. While another paper in this issue applies these methods in the case of acoustic stimuli (Aeschlimann et al., Brain Topogr 2008), we illustrate this approach here in detail for complex visual stimuli.

MSight: an image analysis software for liquid chromatography-mass spectrometry.

Images obtained from high-throughput mass spectrometry (MS) contain information that remains hidden when looking at a single spectrum at a time. Image processing of liquid chromatography-MS datasets can be extremely useful for quality control, experimental monitoring and knowledge extraction. The importance of imaging in differential analysis of proteomic experiments has already been established through two-dimensional gels and can now be foreseen with MS images. We present MSight, a new software designed to construct and manipulate MS images, as well as to facilitate their analysis and comparison.