Extravascular lung water indexed to predicted body weight is a novel predictor of intensive care unit mortality in patients with acute lung injury

Objectives: Acute lung injury and the acute respiratory distress syndrome are characterized by noncardiogenic pulmonary edema, which can be assessed by measurement of extravascular lung water. Traditionally, extravascular lung water has been indexed to actual body weight (mL/kg). Because lung size is dependent on height rather than weight, we hypothesized indexing to predicted body weight may be a better predictor of mortality in acute lung injury/acute respiratory distress syndrome.

A hierarchical multiclass support vector machine incorporated with holistic triple learning units

This paper proposes a new hierarchical learning structure, namely the holistic triple learning (HTL), for extending the binary support vector machine (SVM) to multi-classification problems. For an N-class problem, a HTL constructs a decision tree up to a depth of A leaf node of the decision tree is allowed to be placed with a holistic triple learning unit whose generalisation abilities are assessed and approved. Meanwhile, the remaining nodes in the decision tree each accommodate a standard binary SVM classifier. The holistic triple classifier is a regression model trained on three classes, whose training algorithm is originated from a recently proposed implementation technique, namely the least-squares support vector machine (LS-SVM). A major novelty with the holistic triple classifier is the reduced number of support vectors in the solution. For the resultant HTL-SVM, an upper bound of the generalisation error can be obtained. The time complexity of training the HTL-SVM is analysed, and is shown to be comparable to that of training the one-versus-one (1-vs.-1) SVM, particularly on small-scale datasets. Empirical studies show that the proposed HTL-SVM achieves competitive classification accuracy with a reduced number of support vectors compared to the popular 1-vs-1 alternative.

Translating evidence into practice: A shared priority in public health?

Translational and transdisciplinary research is needed to tackle complex public health problems. This article has three aims. Firstly, to determine how academics and non-academics (practitioners, policy makers and community workers) identified with the goals of the UKCRC Centre of Excellence for Public Health in Northern Ireland and how their attitudes varied in terms of knowledge brokerage and translation. Secondly, to map and analyse the network structure of the public health sector and the placement of the Centre within this. Thirdly, to aggregate responses from members of the network by work setting to construct the trans-sectoral network and devise the Root Mean Sum of Squares to determine the quality and potential value of connections across this network.The analysis was based on data collected from 98 individuals who attended the launch of the Centre in June 2008. Analysis of participant expectations and personal goals suggests that the academic members of the network were more likely to expect the work of the Centre to produce new knowledge than non-academics, but less likely to expect the Centre to generate health interventions and influence health policy. Academics were also less strongly oriented than non-academics to knowledge transfer as a personal goal, though more confident that research findings would be diffused beyond the immediate network. A central core of five nodes is crucial to the overall configuration of the regional public health network in Northern Ireland, with the Centre being well placed to exert influence within this. Though the overall network structure is fairly robust, the connections between some component parts of the network - such as academics and the third sector - are unidirectional.Identifying these differences and core network structure is key to translational and transdisciplinary research. Though exemplified in a regional study, these techniques are generalisable and applicable to many networks of interest: public health, interdisciplinary research or organisational involvement and stakeholder linkage.

Segmentation of optic disc in retinal images using an improved gradient vector flow algorithm

Image segmentation plays an important role in the analysis of retinal images as the extraction of the optic disk provides important cues for accurate diagnosis of various retinopathic diseases. In recent years, gradient vector flow (GVF) based algorithms have been used successfully to successfully segment a variety of medical imagery. However, due to the compromise of internal and external energy forces within the resulting partial differential equations, these methods can lead to less accurate segmentation results in certain cases. In this paper, we propose the use of a new mean shift-based GVF segmentation algorithm that drives the internal/external energies towards the correct direction. The proposed method incorporates a mean shift operation within the standard GVF cost function to arrive at a more accurate segmentation. Experimental results on a large dataset of retinal images demonstrate that the presented method optimally detects the border of the optic disc.

A distributed contention-vector division multiple access (D-CVDMA) protocol for wireless networks

Traditional Time Division Multiple Access (TDMA) protocol provides deterministic periodic collision free data transmissions. However, TDMA lacks flexibility and exhibits low efficiency in dynamic environments such as wireless LANs. On the other hand contention-based MAC protocols such as the IEEE 802.11 DCF are adaptive to network dynamics but are generally inefficient in heavily loaded or large networks. To take advantage of the both types of protocols, a D-CVDMA protocol is proposed. It is based on the k-round elimination contention (k-EC) scheme, which provides fast contention resolution for Wireless LANs. D-CVDMA uses a contention mechanism to achieve TDMA-like collision-free data transmissions, which does not need to reserve time slots for forthcoming transmissions. These features make the D-CVDMA robust and adaptive to network dynamics such as node leaving and joining, changes in packet size and arrival rate, which in turn make it suitable for the delivery of hybrid traffic including multimedia and data content. Analyses and simulations demonstrate that D-CVDMA outperforms the IEEE 802.11 DCF and k-EC in terms of network throughput, delay, jitter, and fairness.