152653 research outputs found
Sort by
Multiple instance cancer detection by boosting regularised trees
Get PDFWe propose a novel multiple instance learning algorithm for cancer detection in histopathology images. With images labelled at image-level, we rst search a set of region-level prototypes by solving a submodular set cover problem. Regularised regression trees are then constructed and combined on the set of prototypes using a multiple instance boosting framework. The method compared favourably with competing methods in experiments on breast cancer tissue microarray images and optical tomographic images of colorectal polyps
Printmaking in the post-print age:critical and creative methods in the context of contemporary art and society
No full textThe collaborative processes between artist/ printmaker and scientists is the subject of my proposed paper and the topic for discussion for the Impact 9 Conference in Hangzhou 2015. In 2006, I undertook a collaborative art-science research project with scientists from the Scottish Crop Research Institute (now the James Hutton Institute), entitled Blueprint for Bacterial Life. Using advanced digital tools I developed the concept of their Genome Diagram into a multimedia interactive installation with animations and music based on the genetic plasticity and evolution of bacterial pathogens. I began by creating a series of etchings and screen-prints using a very subtle range of silvery blues and greys and worked with some specific inks known as interference inks. It was through looking at those prints that the scientists noticed the occurrence of new elements and a very specific event of gene acquisition. By simplifying the diagram into a tonal variation and re-contextualizing the data it revealed information that the scientists had completely overlooked. Their scientific approach to the data was systematic and empirical. By chance, this artistic re-interpretation of the scientific data contributed to a new insight. Rather than simply identifying genes unique to a pathogen, the screen-prints revealed the presence of other genes in all of the bacteria, possibly representing genes essential to all forms of bacteria
Multiple instance cancer detection by boosting regularised trees
Get PDFWe propose a novel multiple instance learning algorithm for cancer detection in histopathology images. With images labelled at image-level, we rst search a set of region-level prototypes by solving a submodular set cover problem. Regularised regression trees are then constructed and combined on the set of prototypes using a multiple instance boosting framework. The method compared favourably with competing methods in experiments on breast cancer tissue microarray images and optical tomographic images of colorectal polyps
Wave transmission in Hostun sand:multiaxial experiments
No full textLaboratory geophysical techniques using waves transmitted by piezoceramic elements have become popular for estimating the very small strain stiffness of soils. Piezoceramic Bender/Extender elements have been installed in a multiaxial cubical cell which is being used to study the multiaxial stiffness of dry Hostun sand. The elements can be used to track wave transmission in both shear and compression modes both across from one side to the opposite side of the cube and diagonally from one side to an adjacent side and thus to deduce the elements of the cross-anisotropic stiffness matrix of the sand. Practicalities of the bender installation reduce the degree of redundancy in these deductions. An assessment of the evolution of elastic anisotropy under axially symmetric stress conditions is presented
Impact of renin-angiotensin system blockade therapy on outcome in aortic stenosis
No full textObjectives The purpose of this study was to investigate the impact of renin-angiotensin system blockade therapy on outcomes in aortic stenosis (AS).Background Angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) are perceived to be relatively contraindicated in AS. However, inhibitors of the renin-angiotensin system may be beneficial in AS through their cardioprotective and beneficial effects on left ventricular remodeling.Methods The Health Informatics dispensed prescribing, morbidity, and mortality database for the population of Tayside, Scotland, was linked through a unique patient identifier to the Tayside echocardiography database (> 110,000 scans). Patients with a diagnosis of AS from 1993 to 2008 were identified. Cox regression model (adjusted for confounding variables) and propensity score analysis were used to assess the impact of ACEIs or ARBs on all-cause mortality and cardiovascular (CV) events (CV death or hospitalizations).Results A total of 2,117 patients with AS (mean age 73 +/- 12 years, 46% men) were identified and 699 (33%) were on ACEI or ARB therapy. Over a mean follow-up of 4.2 years, there were 1,087 (51%) all-cause deaths and 1,018 (48%) CV events. Those treated with ACEIs or ARBs had a significantly lower all-cause mortality with an adjusted hazard ratio of 0.76 (95% confidence interval: 0.62 to 0.92, p < 0.0001) and fewer CV events with an adjusted hazard ratio of 0.77 (95% confidence interval: 0.65 to 0.92, p < 0.0001). The outcome benefits of ACEIs/ARBs were further supported by propensity score analysis.Conclusions This large observational study suggests that ACEI/ARB therapy is associated with an improved survival and a lower risk of CV events in patients with AS. (J Am Coll Cardiol 2011;58:570-6) (C) 2011 by the American College of Cardiology Foundation</p
Allopurinol Benefits Left Ventricular Mass and Endothelial Dysfunction in Chronic Kidney Disease
No full textAllopurinol ameliorates endothelial dysfunction and arterial stiffness among patients without chronic kidney disease (CKD), but it is unknown if it has similar effects among patients with CKD. Furthermore, because arterial stiffness increases left ventricular afterload, any allopurinol-induced improvement in arterial compliance might also regress left ventricular hypertrophy (LVH). We conducted a randomized, double-blind, placebo-controlled, parallel-group study in patients with stage 3 CKD and LVH. We randomly assigned 67 subjects to allopurinol at 300 mg/d or placebo for 9 months; 53 patients completed the study. We measured left ventricular mass index (LVMI) with cardiac magnetic resonance imaging (MRI), assessed endothelial function by flow-mediated dilation (FMD) of the brachial artery, and evaluated central arterial stiffness by pulse-wave analysis. Allopurinol significantly reduced LVH (P = 0.036), improved endothelial function (P = 0.009), and improved the central augmentation index (P = 0.015). This study demonstrates that allopurinol can regress left ventricular mass and improve endothelial function among patients with CKD. Because LVH and endothelial dysfunction associate with prognosis, these results call for further trials to examine whether allopurinol reduces cardiovascular events in patients with CKD and LVH.</p
Optimal feature selection for estimating biomass using a genetic algorithm
No full textWhilst the use of SAR and multispectral image texture has shown promise for estimating tropical forest biomass, there remains uncertainty over the optimum number and type of texture features that provide reliable models to estimate biomass across different forest sites. Previous studies have consistently suggested that selecting an appropriate window size for extracting texture is critical, as small window sizes often exaggerate texture whilst larger window sizes create smoothing effects. There are also a variety of image texture features that have previously been correlated with biomass (e.g. wavelet decomposition and GLCM outputs such as entropy, homogeneity, energy etc.). Neural network regression methods allow any number of these variables (and window sizes) to be included in the regression model. However, this increases the dimensionality of the input data with resultant effects on the ability of the network to learn and generalise. It is desirable to try and restrict the dimensionality of the inputs, particularly when using small samples for predictive modelling, a characteristic of many biomass estimation studies. Common methods for feature selection include Principal Components Analysis (PCA) but are not always suited to this kind of problem. Here, we compare PCA feature selection with a Genetic Algorithm (GA) approach, using separately an artificial neural network and Fuzzy c-means as fitness functions. A combination of texture features were derived from SAR and multispectral images of three tropical forest sites. Evaluation of the optimum combination of these features to estimate aboveground biomass was conducted by applying each of the three feature selection methods in turn, and then using the features selected as inputs to estimate biomass with a neural network. The correlation between the input training data and the unseen testing data was used as a measure of model performance for estimating biomass. The results indicated that features selected using the GA approach with a neural network used as a fitness function produced the strongest relationships with biomass at the three sites (r=0.91, 0.89 and 0.87 for Brazil (n=9), Malaysia (n=9) and Thailand (n=13) respectively), compared to the other GA approach and PCA. In all cases, the texture features and window sizes selected varied, although some commonality in selection between Malaysia and Brazil sites was noted. Overall, the GA approaches selected features that produced stronger relationships than PCA with evidence that these hold much promise for determining the optimum set of inputs for biomass estimation models, although much work is still required. <br/
Data sources:
Access Repository Dashboard
Do you manage University of Dundee Online Publications? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!