Multi-Modal Convolutional Parameterisation Network for Guided Image Inverse Problems.

There are several image inverse tasks, such as inpainting or super-resolution, which can be solved using deep internal learning, a paradigm that involves employing deep neural networks to find a solution by learning from the sample itself rather than a dataset. For example, Deep Image Prior is a technique based on fitting a convolutional neural network to output the known parts of the image (such as non-inpainted regions or a low-resolution version of the image). However, this approach is not well adjusted for samples composed of multiple modalities. In some domains, such as satellite image processing, accommodating multi-modal representations could be beneficial or even essential. In this work, Multi-Modal Convolutional Parameterisation Network (MCPN) is proposed, where a convolutional neural network approximates shared information between multiple modes by combining a core shared network with modality-specific head networks. The results demonstrate that these approaches can significantly outperform the single-mode adoption of a convolutional parameterisation network on guided image inverse problems of inpainting and super-resolution.

Feed Conversion Ratio (FCR) and Performance Group Estimation Based on Predicted Feed Intake for the Optimisation of Beef Production.

This paper reports on the use of estimates of individual animal feed intake (made using time spent feeding measurements) to predict the Feed Conversion Ratio (FCR), a measure of the amount of feed consumed to produce 1 kg of body mass, for an individual animal. Reported research to date has evaluated the ability of statistical methods to predict daily feed intake based on measurements of time spent feeding measured using electronic feeding systems. The study collated data of the time spent eating for 80 beef animals over a 56-day period as the basis for the prediction of feed intake. A Support Vector Regression (SVR) model was trained to predict feed intake and the performance of the approach was quantified. Here, feed intake predictions are used to estimate individual FCR and use this information to categorise animals into three groups based on the estimated Feed Conversion Ratio value. Results provide evidence of the feasibility of utilising the 'time spent eating' data to estimate feed intake and in turn Feed Conversion Ratio (FCR), the latter providing insights that guide farmer decisions on the optimisation of production costs.

A Mapping Review of Real-Time Movement Sonification Systems for Movement Rehabilitation.

Movement sonification is emerging as a useful tool for rehabilitation, with increasing evidence in support of its use. To create such a system requires component considerations outside of typical sonification design choices, such as the dimension of movement to sonify, section of anatomy to track, and methodology of motion capture. This review takes this emerging and highly diverse area of literature and keyword-code existing real-time movement sonification systems, to analyze and highlight current trends in these design choices, as such providing an overview of existing systems. A combination of snowballing through relevant existing reviews and a systematic search of multiple databases were utilized to obtain a list of projects for data extraction. The review categorizes systems into three sections: identifying the link between physical dimension to auditory dimension used in sonification, identifying the target anatomy tracked, identifying the movement tracking system used to monitor the target anatomy. The review proceeds to analyze the systematic mapping of the literature and provide results of the data analysis highlighting common and innovative design choices used, irrespective of application, before discussing the findings in the context of movement rehabilitation. A database containing the mapped keywords assigned to each project are submitted with this review.

Deep Learning for Diabetic Retinopathy Analysis: A Review, Research Challenges, and Future Directions.

Deep learning (DL) enables the creation of computational models comprising multiple processing layers that learn data representations at multiple levels of abstraction. In the recent past, the use of deep learning has been proliferating, yielding promising results in applications across a growing number of fields, most notably in image processing, medical image analysis, data analysis, and bioinformatics. DL algorithms have also had a significant positive impact through yielding improvements in screening, recognition, segmentation, prediction, and classification applications across different domains of healthcare, such as those concerning the abdomen, cardiac, pathology, and retina. Given the extensive body of recent scientific contributions in this discipline, a comprehensive review of deep learning developments in the domain of diabetic retinopathy (DR) analysis, viz., screening, segmentation, prediction, classification, and validation, is presented here. A critical analysis of the relevant reported techniques is carried out, and the associated advantages and limitations highlighted, culminating in the identification of research gaps and future challenges that help to inform the research community to develop more efficient, robust, and accurate DL models for the various challenges in the monitoring and diagnosis of DR.

Behavioural Classification of Cattle Using Neck-Mounted Accelerometer-Equipped Collars.

Monitoring and classification of dairy cattle behaviours is essential for optimising milk yields. Early detection of illness, days before the critical conditions occur, together with automatic detection of the onset of oestrus cycles is crucial for obviating prolonged cattle treatments and improving the pregnancy rates. Accelerometer-based sensor systems are becoming increasingly popular, as they are automatically providing information about key cattle behaviours such as the level of restlessness and the time spent ruminating and eating, proxy measurements that indicate the onset of heat events and overall welfare, at an individual animal level. This paper reports on an approach to the development of algorithms that classify key cattle states based on a systematic dimensionality reduction process through two feature selection techniques. These are based on Mutual Information and Backward Feature Elimination and applied on knowledge-specific and generic time-series extracted from raw accelerometer data. The extracted features are then used to train classification models based on a Hidden Markov Model, Linear Discriminant Analysis and Partial Least Squares Discriminant Analysis. The proposed feature engineering methodology permits model deployment within the computing and memory restrictions imposed by operational settings. The models were based on measurement data from 18 steers, each animal equipped with an accelerometer-based neck-mounted collar and muzzle-mounted halter, the latter providing the truthing data. A total of 42 time-series features were initially extracted and the trade-off between model performance, computational complexity and memory footprint was explored. Results show that the classification model that best balances performance and computation complexity is based on Linear Discriminant Analysis using features selected through Backward Feature Elimination. The final model requires 1.83 ± 1.00 ms to perform feature extraction with 0.05 ± 0.01 ms for inference with an overall balanced accuracy of 0.83.

Developmental differences in the prospective organisation of goal-directed movement between children with autism and typically developing children: A smart tablet serious game study.

Movement is prospective. It structures self-generated engagement with objects and social partners and is fundamental to children's learning and development. In autistic children, previous reports of differences in movement kinematics compared to neurotypical peers suggest that its prospective organisation might be disrupted. Here, we employed a smart tablet serious game paradigm to assess differences in the feedforward and feedback mechanisms of prospective action organisation, between autistic and neurotypical preschool children. We analysed 3926 goal-directed finger movements made during smart-tablet ecological gameplay, from 28 children with Childhood Autism (ICD-10; ASD) and 43 neurotypical children (TD), aged 3-6 years old. Using linear and generalised linear mixed-effect models, we found the ASD group executed movements with longer movement time (MT) and time to peak velocity (TTPV), lower peak velocity (PV), with PV less likely to occur in the first movement unit (MU) and with a greater number of movement units after peak velocity (MU-APV). Interestingly, compared to the TD group, the ASD group showed smaller increases in PV, TTPV and MT with an increase in age (ASD × age interaction), together with a smaller reduction in MU-APV and an increase in MU-APV at shorter target distances (ASD × Dist interaction). Our results are the first to highlight different developmental trends in anticipatory feedforward and compensatory feedback mechanisms of control, contributing to differences in movement kinematics observed between autistic and neurotypical children. These findings point to differences in integration of prospective perceptuomotor information, with implications for embodied cognition and learning from self-generated action in autism.

A Fusion-Based Machine Learning Approach for the Prediction of the Onset of Diabetes.

A growing portfolio of research has been reported on the use of machine learning-based architectures and models in the domain of healthcare. The development of data-driven applications and services for the diagnosis and classification of key illness conditions is challenging owing to issues of low volume, low-quality contextual data for the training, and validation of algorithms, which, in turn, compromises the accuracy of the resultant models. Here, a fusion machine learning approach is presented reporting an improvement in the accuracy of the identification of diabetes and the prediction of the onset of critical events for patients with diabetes (PwD). Globally, the cost of treating diabetes, a prevalent chronic illness condition characterized by high levels of sugar in the bloodstream over long periods, is placing severe demands on health providers and the proposed solution has the potential to support an increase in the rates of survival of PwD through informing on the optimum treatment on an individual patient basis. At the core of the proposed architecture is a fusion of machine learning classifiers (Support Vector Machine and Artificial Neural Network). Results indicate a classification accuracy of 94.67%, exceeding the performance of reported machine learning models for diabetes by ~1.8% over the best reported to date.

Classification of Cattle Behaviours Using Neck-Mounted Accelerometer-Equipped Collars and Convolutional Neural Networks.

Monitoring cattle behaviour is core to the early detection of health and welfare issues and to optimise the fertility of large herds. Accelerometer-based sensor systems that provide activity profiles are now used extensively on commercial farms and have evolved to identify behaviours such as the time spent ruminating and eating at an individual animal level. Acquiring this information at scale is central to informing on-farm management decisions. The paper presents the development of a Convolutional Neural Network (CNN) that classifies cattle behavioural states ('rumination', 'eating' and 'other') using data generated from neck-mounted accelerometer collars. During three farm trials in the United Kingdom (Easter Howgate Farm, Edinburgh, UK), 18 steers were monitored to provide raw acceleration measurements, with ground truth data provided by muzzle-mounted pressure sensor halters. A range of neural network architectures are explored and rigorous hyper-parameter searches are performed to optimise the network. The computational complexity and memory footprint of CNN models are not readily compatible with deployment on low-power processors which are both memory and energy constrained. Thus, progressive reductions of the CNN were executed with minimal loss of performance in order to address the practical implementation challenges, defining the trade-off between model performance versus computation complexity and memory footprint to permit deployment on micro-controller architectures. The proposed methodology achieves a compression of 14.30 compared to the unpruned architecture but is nevertheless able to accurately classify cattle behaviours with an overall F1 score of 0.82 for both FP32 and FP16 precision while achieving a reasonable battery lifetime in excess of 5.7 years.

Composite Laminate Delamination Detection Using Transient Thermal Conduction Profiles and Machine Learning Based Data Analysis.

Delaminations within aerospace composites are of particular concern, presenting within composite laminate structures without visible surface indications. Transmission based thermography techniques using contact temperature sensors and surface mounted heat sources are able to detect reductions in thermal conductivity and in turn impact damage and large disbonds can be detected. However delaminations between Carbon Fibre Reinforced Polymer (CFRP) plies are not immediately discoverable using the technique. The use of transient thermal conduction profiles induced from zonal heating of a CFRP laminate to ascertain inter-laminate differences has been demonstrated and the paper builds on this method further by investigating the impact of inter laminate inclusions, in the form of delaminations, to the transient thermal conduction profile of multi-ply bi-axial CFRP laminates. Results demonstrate that as the distance between centre of the heat source and delamination increase, whilst maintaining the delamination within the heated area, the resultant transient thermal conduction profile is measurably different to that of a homogeneous region at the same distance. The method utilises a supervised Support Vector Classification (SVC) algorithm to detect delaminations using temperature data from either the edge of the defect or the centre during a 140 s ramped heating period to 80 °C. An F1 score in the classification of delaminations or no delamination at an overall accuracy of over 99% in both training and with test data separate from the training process has been achieved using data points effected by transient thermal conduction due to structural dissipation at 56.25 mm.

A Review of Cyber-Ranges and Test-Beds: Current and Future Trends.

Cyber situational awareness has been proven to be of value in forming a comprehensive understanding of threats and vulnerabilities within organisations, as the degree of exposure is governed by the prevailing levels of cyber-hygiene and established processes. A more accurate assessment of the security provision informs on the most vulnerable environments that necessitate more diligent management. The rapid proliferation in the automation of cyber-attacks is reducing the gap between information and operational technologies and the need to review the current levels of robustness against new sophisticated cyber-attacks, trends, technologies and mitigation countermeasures has become pressing. A deeper characterisation is also the basis with which to predict future vulnerabilities in turn guiding the most appropriate deployment technologies. Thus, refreshing established practices and the scope of the training to support the decision making of users and operators. The foundation of the training provision is the use of Cyber-Ranges (CRs) and Test-Beds (TBs), platforms/tools that help inculcate a deeper understanding of the evolution of an attack and the methodology to deploy the most impactful countermeasures to arrest breaches. In this paper, an evaluation of documented CR and TB platforms is evaluated. CRs and TBs are segmented by type, technology, threat scenarios, applications and the scope of attainable training. To enrich the analysis of documented CR and TB research and cap the study, a taxonomy is developed to provide a broader comprehension of the future of CRs and TBs. The taxonomy elaborates on the CRs/TBs dimensions, as well as, highlighting a diminishing differentiation between application areas.

Identifying Defects in Aerospace Composite Sandwich Panels Using High-Definition Distributed Optical Fibre Sensors.

Automated methods for detecting defects within composite materials are highly desirable in the drive to increase throughput, optimise repair program effectiveness and reduce component replacement. Tap-testing has traditionally been used for detecting defects but does not provide quantitative measurements, requiring secondary techniques such as ultrasound to certify components. This paper reports on an evaluation of the use of a distributed temperature measurement system-high-definition fibre optic sensing (HD-FOS)-to identify and characterise crushed core and disbond defects in carbon fibre reinforced polymer (CFRP)-skin, aluminium-core, sandwich panels. The objective is to identify these defects in a sandwich panel by measuring the heat transfer through the panel thickness. A heater mat is used to rapidly increase the temperature of the panel with the HD-FOS sensor positioned on the top surface, measuring temperature. HD-FOS measurements are made using the Luna optical distributed sensor interrogator (ODISI) 9100 system comprising a sensor fabricated using standard single mode fibre (SMF)-20 of external diameter 250 µm, including the cladding. Results show that areas in which defects are present modulate thermal conductivity, resulting in a lower surface temperature. The resultant data are analysed to identify the length, width and type of defect. The non-invasive technique is amenable to application in challenging operational settings, offering high-resolution visualisation and defect classification.

Defect Detection in Aerospace Sandwich Composite Panels Using Conductive Thermography and Contact Sensors.

Sandwich panels consisting of two Carbon Fibre Reinforced Polymer (CFRP) outer skins and an aluminium honeycomb core are a common structure of surfaces on commercial aircraft due to the beneficial strength-weight ratio. Mechanical defects such as a crushed honeycomb core, dis-bonds and delaminations in the outer skins and in the core occur routinely under normal use and are repaired during aerospace Maintenance, Repair and Overhaul (MRO) processes. Current practices rely heavily on manual inspection where it is possible minor defects are not identified prior to primary repair and are only addressed after initial repairs intensify the defects due to thermal expansion during high temperature curing. This paper reports on the development and characterisation of a technique based on conductive thermography implemented using an array of single point temperature sensors mounted on one surface of the panel and the concomitant induced thermal profile generated by a thermal stimulus on the opposing surface to identify such defects. Defects are classified by analysing the differential conduction of thermal energy profiles across the surface of the panel. Results indicate that crushed core and impact damage are detectable using a stepped temperature profile of 80 ∘C The method is amenable to integration within the existing drying cycle stage and reduces the costs of executing the overall process in terms of time-to-repair and manual effort.

The Internet of Things enhancing animal welfare and farm operational efficiency.

The growth in wirelessly enabled sensor network technologies has enabled the low cost deployment of sensor platforms with applications in a range of sectors and communities. In the agricultural domain such sensors have been the foundation for the creation of decision support tools that enhance farm operational efficiency. This Research Reflection illustrates how these advances are assisting dairy farmers to optimise performance and illustrates where emerging sensor technology can offer additional benefits. One of the early applications for sensor technology at an individual animal level was the accurate identification of cattle entering into heat (oestrus) to increase the rate of successful pregnancies and thus optimise milk yield per animal. This was achieved through the use of activity monitoring collars and leg tags. Additional information relating to the behaviour of the cattle, namely the time spent eating and ruminating, was subsequently derived from collars giving further insights of economic value into the wellbeing of the animal, thus an enhanced range of welfare related services have been provisioned. The integration of the information from neck-mounted collars with the compositional analysis data of milk measured at a robotic milking station facilitates the early diagnosis of specific illnesses such as mastitis. The combination of different data streams also serves to eliminate the generation of false alarms, improving the decision making capability. The principle of integrating more data streams from deployed on-farm systems, for example, with feed composition data measured at the point of delivery using instrumented feeding wagons, supports the optimisation of feeding strategies and identification of the most productive animals. Optimised feeding strategies reduce operational costs and minimise waste whilst ensuring high welfare standards. These IoT-inspired solutions, made possible through Internet-enabled cloud data exchange, have the potential to make a major impact within farming practices. This paper gives illustrative examples and considers where new sensor technology from the automotive industry may also have a role.

Non-Destructive Identification of Fibre Orientation in Multi-Ply Biaxial Laminates Using Contact Temperature Sensors.

Fibre orientation within composite structures dictates the material properties of the laminate once cured. The ability to accurately and automatically assess fibre orientation of composite parts is a significant enabler in the goal to optimise the established processes within aftermarket aerospace industries. Incorrect ply lay-up results in a structure with undesirable material properties and as such, has the potential to fail under safe working loads. Since it is necessary to assure structural integrity during re-manufacture and repair assessment, the paper demonstrates a novel method of readily and non-destructively determining fibre orientation throughout multi-ply Biaxial woven composite laminates using point temperature contact sensors and data analysis techniques. Once cured, only the outermost laminates are visible to assess orientation. The inspection method is conducted visually, with reference guides to allow for rapid adoption with minimum training, as well as harnessing established temperature sensors within the Maintenance Repair and Overhaul (MRO) environment. The system is amenable to integration within existing repair/re-manufacture processes without significant impact to process flow. The method is able to identify noisy samples with an accuracy, precision and recall of 0.9, and for synthetically created samples of double the cure ply thickness, a precision of 0.75, a recall of 0.7 and an accuracy of 0.87.

Automatic Annotation of Subsea Pipelines using Deep Learning.

Regulatory requirements for sub-sea oil and gas operators mandates the frequent inspection of pipeline assets to ensure that their degradation and damage are maintained at acceptable levels. The inspection process is usually sub-contracted to surveyors who utilize sub-sea Remotely Operated Vehicles (ROVs), launched from a surface vessel and piloted over the pipeline. ROVs capture data from various sensors/instruments which are subsequently reviewed and interpreted by human operators, creating a log of event annotations; a slow, labor-intensive and costly process. The paper presents an automatic image annotation framework that identifies/classifies key events of interest in the video footage viz. exposure, burial, field joints, anodes, and free spans. The reported methodology utilizes transfer learning with a Deep Convolutional Neural Network (ResNet-50), fine-tuned on real-life, representative data from challenging sub-sea environments with low lighting conditions, sand agitation, sea-life and vegetation. The network outputs are configured to perform multi-label image classifications for critical events. The annotation performance varies between 95.1% and 99.7% in terms of accuracy and 90.4% and 99.4% in terms of F1-Score depending on event type. The performance results are on a per-frame basis and corroborate the potential of the algorithm to be the foundation for an intelligent decision support framework that automates the annotation process. The solution can execute annotations in real-time and is significantly more cost-effective than human-only approaches.

Identification of the Rumination in Cattle Using Support Vector Machines with Motion-Sensitive Bolus Sensors.

The reticuloruminal function is central to the digestive efficiency in ruminants. For cattle, collar- and ear tag-based accelerometer monitors have been developed to assess the time spent ruminating on an individual animal. Cattle that are ill feed less and so ruminate less, thus, the estimation of the time spent ruminating provides insights into the health of individual animals. pH boluses directly provide information on the reticuloruminal function within the rumen and extended (three hours or more) periods during which the ruminal pH value remains below 5.6 is an indicator that dysfunction and poor welfare are likely. Accelerometers, incorporated into the pH boluses, have been used to indicate changes in behaviour patterns (high/low activity), utilised to detect the onset of oestrus. The paper demonstrates for the first time that by processing the reticuloruminal motion, it is possible to recover rumination periods. Reticuloruminal motion energy and the time between reticuloruminal contractions are used as inputs to a Support Vector Machine (SVM) to identify rumination periods with an overall accuracy of 86.1%, corroborated by neck mounted rumination collars.

Evaluation of reticuloruminal pH measurements from individual cattle: Sampling strategies for the assessment of herd status.

The application of pH observations to clinical practice in dairy cattle is based on criteria derived primarily from single time-point observations more than 20 years ago. The aims of this study were to evaluate these criteria using data collected using continuous recording methods; to make recommendations that might improve their interpretation; and to determine the relationship between the number of devices deployed in a herd and the accuracy of the resulting estimate of the herd-mean reticuloruminal pH. The study made use of 815,475 observations of reticuloruminal pH values obtained from 75 cattle in three herds (one beef and two twice-daily milking herds) to assess sampling strategies for the diagnosis of sub-acute rumen acidosis (SARA), and to evaluate the ability of different numbers of bolus devices to accurately estimate the true herd-mean reticuloruminal pH value at any time. The traditional criteria for SARA provide low diagnostic utility, the probability of detection of animals with pH values below specified thresholds being affected by a strong effect of time of day and herd. The analysis suggests that regardless of time of feeding, sampling should be carried out in the late afternoon or evening to obtain a reasonable probability of detection of animals with pH values below the threshold level. The among-cow variation varied strongly between herds, but for a typical herd, if using reticuloruminal pH boluses to detect a predisposition to fermentation disorders while feeding a diet that is high in rapidly fermentable carbohydrates, it is recommended to use a minimum of nine boluses.

Crystal Shape Modification via Cycles of Growth and Dissolution in a Tubular Crystallizer.

Besides size and polymorphic form, crystal shape takes a central role in engineering advanced solid materials for the pharmaceutical and chemical industries. This work demonstrates how multiple cycles of growth and dissolution can manipulate the habit of an acetylsalicylic acid crystal population. Considerable changes of the crystal habit could be achieved within minutes due to rapid cycling, i.e., up to 25 cycles within <10 min. The required fast heating and cooling rates were facilitated using a tubular reactor design allowing for superior temperature control. The face-specific interactions between solvent and the crystals' surface result in face-specific growth and dissolution rates and hence alterations of the final shape of the crystals in solution. Accurate quantification of the crystal shapes was essential for this work, but is everything except simple. A commercial size and shape analyzer had to be adapted to achieve the required accuracy. Online size, and most important shape, analysis was achieved using an automated microscope equipped with a flow-through cell, in combination with a dedicated image analysis routine for particle tracking and shape analysis. Due to the implementation of this analyzer, capable of obtaining statistics on the crystals' shape while still in solution (no sampling and manipulation required), the dynamic behavior of the size shape distribution could be studied. This enabled a detailed analysis of the solvent's effect on the change in crystal habit.