Search

Breast cancer identification is a arduous process and diagnosing it using Haematoxylin and Eosin (H&E) stained pathology images is a significant challenge, with pathologists struggling to segment cancer nuclei accurately. This study will evaluate the efficacy of different methods utilising deep learning techniques for breast cancer nuclei segmentation, with a particular emphasis on U-Net architectures. The proposed methodology is divided into four stages: image enhancement, individual nuclei segmentation, feature extraction, and whole image binary segmentation. This work then conducts a rigorous comparison of different segmentation techniques.

Over the last decades, sleep apnea has become one of the most prevalent healthcare problems. Diagnosis and treatment monitoring are key elements when it comes to addressing this public health crisis. A problem for diagnosis and treatment monitoring is a chronic lack of specialized lab facilities which results in long waiting times or the absence of such services. This can delay appropriate treatment which might prolong living with sleep apnea and thereby leading to health issues due to poor sleep. We address this problem with a smart sleep apnea detection service based on Heart Rate Variably (HRV) analysis. The service incorporates Internet of Medical Things (IoMT), mobile technology (MT), and advanced Artificial Intelligence (AI). The measured signals are relayed by a smart phone into a cloud server via IoMT protocols. Once the data is stored in the cloud server, a deep learning (DL) algorithm is used to detect sleep apnea events. Detecting these events can trigger a warning message which is sent to care givers. The smart sleep apnea detection service is beneficial for patients who find it difficult to access specialized lab facilities for diagnosis or treatment monitoring. Furthermore, the system prolongs the observation period, which can improve the diagnosis accuracy. The resource requirements for the proposed service are lower when compared to clinical facilities, this might lead to significant cost savings for healthcare providers.

Real-time acquisition of large amounts of machine operating data is now increasingly common due to recent advances in Industry 4.0 technologies. A key benefit to factory operators of this large scale data acquisition is in the ability to perform real-time condition monitoring and early-stage fault detection and diagnosis on industrial machinery&mdash, with the potential to reduce machine down-time and thus operating costs. The main contribution of this work is the development of an intelligent fault diagnosis method capable of operating on these real-time data streams to provide early detection of developing problems under variable operating conditions. We propose a novel dual-path recurrent neural network with a wide first kernel and deep convolutional neural network pathway (RNN-WDCNN) capable of operating on raw temporal signals such as vibration data to diagnose rolling element bearing faults in data acquired from electromechanical drive systems. RNN-WDCNN combines elements of recurrent neural networks (RNNs) and convolutional neural networks (CNNs) to capture distant dependencies in time series data and suppress high-frequency noise in the input signals. Experimental results on the benchmark Case Western Reserve University (CWRU) bearing fault dataset show RNN-WDCNN outperforms current state-of-the-art methods in both domain adaptation and noise rejection tasks.

The optimisation and validation of a classifiers performance when applied to real\ud world problems is not always effectively shown. In much of the literature describing\ud the application of artificial neural network architectures to Human Activity\ud Recognition (HAR) problems, postural transitions are grouped together and treated as\ud a singular class. This paper proposes, investigates and validates the development of\ud an optimised artificial neural network based on Long-Short Term Memory techniques\ud (LSTM), with repeated cross validation used to validate the performance of the\ud classifier. The results of the optimised LSTM classifier are comparable or better to\ud that of previous research making use of the same dataset, achieving 95% accuracy\ud under repeated 10-fold cross validation using grouped postural transitions. The work\ud in this paper also achieves 94% accuracy under repeated 10-fold cross validation\ud whilst treating each common postural transition as a separate class (and thus\ud providing more context to each activity).

The distribution of computational tasks across a diverse set of geographically distributed heterogeneous resources is a critical issue in the realisation of true computational grids. Conventionally, workload allocation algorithms are divided into static and dynamic approaches. Whilst dynamic approaches frequently outperform static schemes, they usually require the collection and processing of detailed system information at frequent intervals - a task that can be both time consuming and unreliable in the real-world. This paper introduces a novel workload allocation algorithm for optimally distributing the workload produced by the arrival of batches of jobs. Results show that, for the arrival of batches of jobs, this workload allocation algorithm outperforms other commonly used algorithms in the static case. A hybrid scheduling approach (using this workload allocation algorithm), where information about the speed of computational resources is inferred from previously completed jobs, is then introduced and the efficiency of this approach demonstrated using a real world computational grid. These results are compared to the same workload allocation algorithm used in the static case and it can be seen that this hybrid approach comprehensively outperforms the static approach.

Introduction: Over the last 25 years there has been significant work carried out in producing risk prediction models for patients admitted to critical care units. The most recent of these models is the Intensive Care National Audit and Research Centre (ICNARC) model developed in 2007 (1) which uses data from 231,930 admissions to 163 critical care units to develop and validate a UK based model outperforming other approaches (with an average c index of 0.863). \ud Aims: This research aims to present an artificial neural network based model for critical care admissions that improves over the ICNARC model in terms of the discrimination across the data set used in this study. \ud Results: Figure 1 shows a comparison between the receiver operator characteristics (ROC) curve for our artificial neural network (ANN) model and the ICNARC model presented in (1). This figure shows the ROC curve and point-wise confidence intervals for the true positive values of both our model (in blue) and the ICNARC model (in red).\ud In comparison, our artificial neural network classification model produces an average c value of 0.8983 in 10 fold cross validation of our data compared to a c value of 0.8306 for the ICNARC model using the same data set (consisting of 642 patients admitted to North Middlesex Hospital critical care unit over a 28 month period. Data excludes 432 patients where data was incomplete).\ud Conclusion: Our classification model provides a percentage risk score that outperforms the ICNARC model. \ud This classification model does suffer from some of same issues surrounding the ICNARC model – for instance, the influence of some of the parameters within both models can be unclear to clinicians trying to predict the survival of individual patients. However, further work is ongoing to improve the transparency of this model

This preliminary study aims to record the brainwaves of two experienced archers, whist undertaking the process of aiming and shooting arrows at a target. The brainwaves are then analysed for repeatability and dominant characteristics within individual EEG activity. Images of the archers are also recorded to establish reference points within the shot cycle for correlating the EEG data sets.

Modelling, Optimisation and Decision Support tools are vital in most areas of engineering. As part of the Distributed Aircraft Maintenance Environment (DAME) e-Science project, a virtual “work-bench” has been developed that can aid the engineer in solving engineering design problems. While the problem that motivated the development of these tools was taken from the aerospace industry, this sort of approach has broader application in areas such as automotive and marine engineering, as well as in the medical industry.