• A deep learning approach to building an intelligent video surveillance system

      Xu, Jie; orcid: 0000-0002-2287-9971; email: jie.xu-4@student.manchester.ac.uk (Springer US, 2020-10-07)
      Abstract: Recent advances in the field of object detection and face recognition have made it possible to develop practical video surveillance systems with embedded object detection and face recognition functionalities that are accurate and fast enough for commercial uses. In this paper, we compare some of the latest approaches to object detection and face recognition and provide reasons why they may or may not be amongst the best to be used in video surveillance applications in terms of both accuracy and speed. It is discovered that Faster R-CNN with Inception ResNet V2 is able to achieve some of the best accuracies while maintaining real-time rates. Single Shot Detector (SSD) with MobileNet, on the other hand, is incredibly fast and still accurate enough for most applications. As for face recognition, FaceNet with Multi-task Cascaded Convolutional Networks (MTCNN) achieves higher accuracy than advances such as DeepFace and DeepID2+ while being faster. An end-to-end video surveillance system is also proposed which could be used as a starting point for more complex systems. Various experiments have also been attempted on trained models with observations explained in detail. We finish by discussing video object detection and video salient object detection approaches which could potentially be used as future improvements to the proposed system.
    • A deep-learning model for urban traffic flow prediction with traffic events mined from twitter

      Essien, Aniekan; orcid: 0000-0001-9501-0647; email: a.e.essien@swansea.ac.uk; email: aniekan.essien@manchester.ac.uk; Petrounias, Ilias; Sampaio, Pedro; Sampaio, Sandra (Springer US, 2020-03-14)
      Abstract: Short-term traffic parameter forecasting is critical to modern urban traffic management and control systems. Predictive accuracy in data-driven traffic models is reduced when exposed to non-recurring or non-routine traffic events, such as accidents, road closures, and extreme weather conditions. The analytical mining of data from social networks – specifically twitter – can improve urban traffic parameter prediction by complementing traffic data with data representing events capable of disrupting regular traffic patterns reported in social media posts. This paper proposes a deep learning urban traffic prediction model that combines information extracted from tweet messages with traffic and weather information. The predictive model adopts a deep Bi-directional Long Short-Term Memory (LSTM) stacked autoencoder (SAE) architecture for multi-step traffic flow prediction trained using tweets, traffic and weather datasets. The model is evaluated on an urban road network in Greater Manchester, United Kingdom. The findings from extensive empirical analysis using real-world data demonstrate the effectiveness of the approach in improving prediction accuracy when compared to other classical/statistical and machine learning (ML) state-of-the-art models. The improvement in predictive accuracy can lead to reduced frustration for road users, cost savings for businesses, and less harm to the environment.
    • A Delphi Study Investigating Clinicians’ Views on Access to, Delivery of, and Adaptations of MBCT in the UK Clinical Settings

      Williams, Kate; orcid: 0000-0001-8167-0951; email: kate.williams-4@manchester.ac.uk; Hartley, Samantha; Taylor, Peter (Springer US, 2021-07-30)
      Abstract: Objectives: Mindfulness-based cognitive therapy (MBCT) is a well-evidenced relapse-prevention intervention for depression with a growing evidence-base for use in other clinical populations. The UK initiatives have outlined plans for increasing access to MBCT in clinical settings, although evidence suggests that access remains limited. Given the increased popularity and access to MBCT, there may be deviations from the evidence-base and potential risks of harm. We aimed to understand what clinicians believe should be best clinical practice regarding access to, delivery of, and adaptations to MBCT. Methods: We employed a two-stage Delphi methodology. First, to develop statements around best practices, we consulted five mindfulness-based experts and reviewed the literature. Second, a total of 59 statements were taken forward into three survey rating rounds. Results: Twenty-nine clinicians completed round one, with 25 subsequently completing both rounds two and three. Forty-four statements reached consensus; 15 statements did not. Clinicians agreed with statements regarding sufficient preparation for accessing MBCT, adherence to the evidence-base and good practice guidelines, consideration of risks, sufficient access to training, support, and resources within services, and carefully considered adaptations. The consensus was not reached on statements which reflected a lack of evidence-base for specific clinical populations or the complex decision-making processes involved in delivering and making adaptations to MBCT. Conclusions: Our findings highlight the delicate balance of maintaining a client-centred and transparent approach whilst adhering to the evidence-base in clinical decisions around access to, delivery of, and adaptations in MBCT and have important wide-reaching implications.
    • A Lucas–Lehmer approach to generalised Lebesgue–Ramanujan–Nagell equations

      Patel, Vandita; orcid: 0000-0003-0252-963X; email: vandita.patel@manchester.ac.uk (Springer US, 2021-06-10)
      Abstract: We describe a computationally efficient approach to resolving equations of the form C1x2+C2=yn in coprime integers, for fixed values of C1, C2 subject to further conditions. We make use of a factorisation argument and the Primitive Divisor Theorem due to Bilu, Hanrot and Voutier.
    • A Multi-Model Pipeline for Translational Intracerebral Haemorrhage Research

      Withers, Sarah E.; Parry-Jones, Adrian R.; Allan, Stuart M.; Kasher, Paul R.; orcid: 0000-0002-9213-502X; email: paul.kasher@manchester.ac.uk (Springer US, 2020-07-07)
      Abstract: Apart from acute and chronic blood pressure lowering, we have no specific medications to prevent intracerebral haemorrhage (ICH) or improve outcomes once bleeding has occurred. One reason for this may be related to particular limitations associated with the current pre-clinical models of ICH, leading to a failure to translate into the clinic. It would seem that a breakdown in the ‘drug development pipeline’ currently exists for translational ICH research which needs to be urgently addressed. Here, we review the most commonly used pre-clinical models of ICH and discuss their advantages and disadvantages in the context of translational studies. We propose that to increase our chances of successfully identifying new therapeutics for ICH, a bi-directional, 2- or 3-pronged approach using more than one model species/system could be useful for confirming key pre-clinical observations. Furthermore, we highlight that post-mortem/ex-vivo ICH patient material is a precious and underused resource which could play an essential role in the verification of experimental results prior to consideration for further clinical investigation. Embracing multidisciplinary collaboration between pre-clinical and clinical ICH research groups will be essential to ensure the success of this type of approach in the future.
    • A New Argument for No-Fault Compensation in Health Care: The Introduction of Artificial Intelligence Systems

      Holm, Søren; Stanton, Catherine; orcid: 0000-0003-0405-0665; email: catherine.stanton@manchester.ac.uk; Bartlett, Benjamin (Springer US, 2021-03-21)
      Abstract: Artificial intelligence (AI) systems advising healthcare professionals will be widely introduced into healthcare settings within the next 5–10 years. This paper considers how this will sit with tort/negligence based legal approaches to compensation for medical error. It argues that the introduction of AI systems will provide an additional argument pointing towards no-fault compensation as the better legal solution to compensation for medical error in modern health care systems. The paper falls into four parts. The first part rehearses the main arguments for and against no-fault compensation. The second explains why it is likely that AI systems will be widely introduced. The third part analyses why it is difficult to fit AI systems into fault-based compensation systems while the final part suggests how no-fault compensation could provide a possible solution to such challenges.
    • A novel and potentially scalable CVD-based route towards SnO 2:Mo thin films as transparent conducting oxides

      Ma, Tianlei; Nikiel, Marek; Thomas, Andrew G.; Missous, Mohamed; Lewis, David J.; orcid: 0000-0001-5950-1350; email: david.lewis-4@manchester.ac.uk (Springer US, 2021-07-12)
      Abstract: In this report, we prepared transparent and conducting undoped and molybdenum-doped tin oxide (Mo–SnO2) thin films by aerosol-assisted chemical vapour deposition (AACVD). The relationship between the precursor concentration in the feed and in the resulting films was studied by energy-dispersive X-ray spectroscopy, suggesting that the efficiency of doping is quantitative and that this method could potentially impart exquisite control over dopant levels. All SnO2 films were in tetragonal structure as confirmed by powder X-ray diffraction measurements. X-ray photoelectron spectroscopy characterisation indicated for the first time that Mo ions were in mixed valence states of Mo(VI) and Mo(V) on the surface. Incorporation of Mo6+ resulted in the lowest resistivity of 7.3×10-3Ωcm, compared to pure SnO2 films with resistivities of 4.30×10-2Ωcm. Meanwhile, a high transmittance of 83% in the visible light range was also acquired. This work presents a comprehensive investigation into impact of Mo doping on SnO2 films synthesised by AACVD for the first time and establishes the potential for scalable deposition of SnO2:Mo thin films in TCO manufacturing. Graphical abstract:
    • A shift in Clinical Learning Modalities—Telemedicine Training in UK Undergraduate Medical Students

      Tang, Celion; orcid: 0000-0002-5054-1425; email: celion.tang@student.manchester.ac.uk; Klair, Anjeevan Kaur; Hussain, Amir (Springer US, 2021-01-09)
    • ABBA: adaptive Brownian bridge-based symbolic aggregation of time series

      Elsworth, Steven; Güttel, Stefan; orcid: 0000-0003-1494-4478; email: stefan.guettel@manchester.ac.uk (Springer US, 2020-06-03)
      Abstract: A new symbolic representation of time series, called ABBA, is introduced. It is based on an adaptive polygonal chain approximation of the time series into a sequence of tuples, followed by a mean-based clustering to obtain the symbolic representation. We show that the reconstruction error of this representation can be modelled as a random walk with pinned start and end points, a so-called Brownian bridge. This insight allows us to make ABBA essentially parameter-free, except for the approximation tolerance which must be chosen. Extensive comparisons with the SAX and 1d-SAX representations are included in the form of performance profiles, showing that ABBA is often able to better preserve the essential shape information of time series compared to other approaches, in particular when time warping measures are used. Advantages and applications of ABBA are discussed, including its in-built differencing property and use for anomaly detection, and Python implementations provided.
    • Advances in Microstructural Understanding of Wrought Aluminum Alloys

      Robson, J. D.; email: joseph.d.robson@manchester.ac.uk; Engler, O.; Sigli, C.; Deschamps, A.; Poole, W. J. (Springer US, 2020-07-08)
      Abstract: Wrought aluminum alloys are an attractive option in the quest for lightweight, recyclable, structural materials. Modern wrought aluminum alloys depend on control of complex microstructures to obtain their properties. This requires an understanding of the coupling between alloy composition, processing, and microstructure. This paper summarizes recent work to understand microstructural evolution in such alloys, utilizing the advanced characterization techniques now available such as atom probe tomography, high-resolution electron microscopy, and synchrotron X-ray diffraction and scattering. New insights into precipitation processes, deformation behavior, and texture evolution are discussed. Recent progress in predicting microstructural evolution using computer modeling is also summarized.
    • Application of Imaging Techniques to Determine the Post-Yield Behaviour of the Heterogeneous Microstructure of Friction Stir Welds

      Ramachandran, S.; orcid: 0000-0002-6881-2940; email: s.ramachandran@manchester.ac.uk; Lakshminarayanan, A. K.; Reed, P. A. S.; Dulieu-Barton, J. M. (Springer US, 2021-04-21)
      Abstract: Background: Friction Stir Welding (FSW) causes intense plastic deformation and consequent thermomechanical interactions resulting in a localised heterogeneous microstructure. To understand the weld mechanical behaviour, it is necessary to identify each microstructural sub-region in the weld. Objective: Determine the relationship between the local microstructure and mechanical behaviour of the different microstructural regions in a FSW. Methods: Scanning electron microscopy (SEM) identified the microstructural sub-regions of an FSW joint. A novel High-Resolution Digital Image Correlation (HR-DIC) methodology enabled the determination of full-field strain response to provide the mechanical behaviour of the FSW sub-regions. X-ray computed tomography (CT) identified the geometry of the FSW and material composition. Results: The grain morphology in the FSW varied in the stir zone with a fine grain structure in the weld nugget and larger grains in the thermomechanical affected zone (TMAZ); the grains were larger in the retreating side (RS) compared to the advancing side (AS). Tungsten deposits were found in the weld nugget and attributed to tool wear. The mechanical properties of the weld subregions showed that the material in the stir zone had a greater yield strength than the base material and the RS of the FSW was much more ductile than the weld nugget and the AS side. The tungsten distributions in the stir zone correlated with the local mechanical behaviour. Conclusions: A novel methodology is developed that combines microstructural observations with HR-DIC enabling, for the first time, the FSW sub-region mechanical behaviour, to be related to the local grain morphology and inclusions caused by tool wear.
    • Bulk and Confined Benzene-Cyclohexane Mixtures Studied by an Integrated Total Neutron Scattering and NMR Method

      Hughes, Terri-Louise; orcid: 0000-0002-6389-8669; email: terri-louise.hughes@manchester.ac.uk; Falkowska, Marta; orcid: 0000-0003-0888-005X; Leutzsch, Markus; orcid: 0000-0001-8171-9399; Sederman, Andrew J.; orcid: 0000-0002-7866-5550; Mantle, Mick D.; Headen, Thomas F.; orcid: 0000-0003-0095-5731; Youngs, Tristan G. A.; orcid: 0000-0003-3538-5572; Bowron, Daniel T.; orcid: 0000-0002-4557-1929; Hardacre, Christopher; orcid: 0000-0001-7256-6765 (Springer US, 2021-04-23)
      Abstract: Herein mixtures of cyclohexane and benzene have been investigated in both the bulk liquid phase and when confined in MCM-41 mesopores. The bulk mixtures have been studied using total neutron scattering (TNS), and the confined mixtures have been studied by a new flow-utilising, integrated TNS and NMR system (Flow NeuNMR), all systems have been analysed using empirical potential structure refinement (EPSR). The Flow NeuNMR setup provided precise time-resolved chemical sample composition through NMR, overcoming the difficulties of ensuring compositional consistency for computational simulation of data ordinarily found in TNS experiments of changing chemical composition—such as chemical reactions. Unique to the liquid mixtures, perpendicularly oriented benzene molecules have been found at short distances from the cyclohexane rings in the regions perpendicular to the carbon–carbon bonds. Upon confinement of the hydrocarbon mixtures, a stronger parallel orientational preference of unlike molecular dimers, at short distances, has been found. At longer first coordination shell distances, the like benzene molecular spatial organisation within the mixture has also found to be altered upon confinement.
    • Children’s Narratives of Family Life in Ghana: A Cultural Lens via Story Stems

      Hosny, Nadine M.; Danquah, Adam; Berry, Katherine; Wan, Ming Wai; email: M.W.Wan@manchester.ac.uk (Springer US, 2020-10-12)
      Abstract: A Western worldview pervades the social and psychological study of children. The current study employed a story-stem method to qualitatively explore the daily and family life experiences of young Ghanaian primary school children in urban Ghana through their story narratives. The recorded narratives of 69 5- to 8-year-old children were elicited through presenting stems of common child-caregiver scenarios and thematically analysed. Five overarching themes were identified: daily routines and concerns, child-caregiver interactions, spirituality, death-related fears and depictions, and responses to injury. The narrative themes reveal the likely mental and physical occupations of these children, as characterized by household chores, financial concerns and school concerns, as well as underlying cultural values through their portrayals of parenting values, parental discipline and spiritual beliefs. Story stems depicting common childhood problems (e.g. knee injury) sometimes evoked death and hospitalization subthemes following an escalation of health problems. Possible interpretations for these findings are discussed, taking into account cultural, developmental and emotional factors. Story stems are a promising and developmentally appropriate tool for qualitative analysts to investigate the experiences and worldview of young children in non-Western cultures.
    • Composite matrices from group rings, composite G -codes and constructions of self-dual codes

      Dougherty, Steven T.; Gildea, Joe; Korban, Adrian; orcid: 0000-0001-5206-6480; email: adrian3@windowslive.com; Kaya, Abidin (Springer US, 2021-05-19)
      Abstract: In this work, we define composite matrices which are derived from group rings. We extend the idea of G-codes to composite G-codes. We show that these codes are ideals in a group ring, where the ring is a finite commutative Frobenius ring and G is an arbitrary finite group. We prove that the dual of a composite G-code is also a composite G-code. We also define quasi-composite G-codes. Additionally, we study generator matrices, which consist of the identity matrices and the composite matrices. Together with the generator matrices, the well known extension method, the neighbour method and its generalization, we find extremal binary self-dual codes of length 68 with new weight enumerators for the rare parameters γ=7, 8 and 9. In particular, we find 49 new such codes. Moreover, we show that the codes we find are inaccessible from other construction
    • Coping Style and Early Life Vocalizations in the Common Marmoset ( Callithrix jacchus )

      Díaz, Sergio; email: s.gonzalezdiaz@chester.ac.uk; Sánchez, Susana; Fidalgo, Ana (Springer US, 2020-05-20)
      Abstract: Coping styles describe behavioral differences during stressful or challenging situations. Coping styles are stable over time but little is known about early life manifestation and development of these behavioral differences. We aimed to investigate if differences in the way marmosets produce vocalizations at an early age are related to their coping style in the future. We studied 14 common marmosets (Callithrix jacchus) from three social groups housed at the marmoset colony at Universidad Autónoma de Madrid. We recorded the vocalizations of each marmoset in isolation at 15–17 days of age, analyzing latency to vocalize and calling rate of phee and tsik calls. To measure coping style, we introduced a novel stimulus to the group cages when infants were 3 months old and recorded exploration, headcocking, and approaches to the stimulus. The results showed negative relationships between the latency of phee call (a long-range contact call) at 15–17 days and frequency of exploration and approach to the novel stimulus at 3 months, although both correlations fall above the cut-off points for the false discovery rate. Marmosets that gave long-range calls sooner at 15–17 days of age also showed more exploratory behaviors at 3 months. The results also showed group differences in exploration at 3 months, and twins were more similar to each other than to other infants in the sample. There were no group differences in early vocalizations and no sex differences in any variable. These findings suggest that coping style is stable from as early as 15–17 days after birth and suggest that the group can influence exploration in marmosets.
    • Deformation Enhanced Diffusion in Aluminium Alloys

      Robson, J. D.; email: joseph.robson@manchester.ac.uk (Springer US, 2020-08-18)
      Abstract: Deformation introduces defects such as dislocations and excess vacancies that can strongly influence diffusion rates in aluminium alloys. This is of great importance in understanding the effect of deformation on processes such as precipitation hardening, which can be accelerated by orders of magnitude by the defects introduced. In this work, a simple and widely used classical model has been employed to explore the effect of process variables on diffusion enhancement due to deformation-induced excess vacancies and dislocations. It is demonstrated that in aluminium alloys, the strain rates and temperatures used in processing and testing encompass a range of regimes. At low strain rate or high temperature, a steady state becomes established in which the deformation enhancement depends on strain rate but not strain. Conversely, at high strain rate or low temperature, deformation enhancement is insensitive to strain rate but increases with strain. For all conditions, the effect of excess vacancies is much stronger than the direct effect of pipe diffusion along dislocations. The predicted deformation-induced vacancy concentration can exceed that expected after rapid quenching by an order of magnitude, but the lifetime of the deformation-induced vacancies is much shorter. Finally, limitations of the classical model and suggestions for improvement are proposed.
    • Enabling Dynamic System Integration on Maxeler HLS Platforms

      Kritikakis, Charalampos; email: charalampos.kritikakis@manchester.ac.uk; Koch, Dirk (Springer US, 2020-08-09)
      Abstract: High Level Synthesis (HLS) tools enable application domain experts to implement applications and algorithms on FPGAs. The majority of present FPGA applications is following a stream processing model which is almost entirely implemented statically and not exploiting the benefits enabled by partial reconfiguration. In this paper, we propose a generic approach for implementing and using partial reconfiguration through an HLS design flow for Maxeler platforms. Our flow extracts HLS generated HDL code from the Maxeler compilation process in order to implement a static FPGA infrastructure as well as run-time reconfigurable stream processing modules. As a distinct feature, our infrastructure can accommodate multiple partial modules in a pipeline daisy-chained manner, which aligns directly to Maxeler’s dataflow programming paradigm. The benefits of the proposed flow are demonstrated by a case study of a dynamically reconfigurable video processing pipeline delivering 6.4GB/s throughput.
    • Ensuring Risk Awareness of Vulnerable Patients in the Post- Montgomery Era: Treading a Fine Line

      Talukdar, Sandip; email: sandip.talukdar@gmail.com; email: sandip.talukdar@postgrad.manchester.ac.uk (Springer US, 2020-05-17)
      Abstract: The 2015 UK Supreme Court judgment in Montgomery v Lanarkshire reinforces the importance of informed consent to medical treatment. This paper suggests that Montgomery recognises the challenge faced by vulnerable individuals in choosing between treatment options and making decisions with appreciation of information about material risks. The judgment endorses a form of weak paternalism to safeguard such persons, which is not disrespectful of the aggregate principles of the Mental Capacity Act 2005. But ethical practice requires professionals to tread carefully between weak and hard paternalism in the context of therapeutic interactions with vulnerable patients, while ensuring their awareness of material risks.
    • Error Estimates of a Continuous Galerkin Time Stepping Method for Subdiffusion Problem

      Yan, Yuyuan; Egwu, Bernard A.; Liang, Zongqi; Yan, Yubin; orcid: 0000-0002-5686-5017; email: y.yan@chester.ac.uk (Springer US, 2021-07-29)
      Abstract: A continuous Galerkin time stepping method is introduced and analyzed for subdiffusion problem in an abstract setting. The approximate solution will be sought as a continuous piecewise linear function in time t and the test space is based on the discontinuous piecewise constant functions. We prove that the proposed time stepping method has the convergence order O(τ1+α), α∈(0, 1) for general sectorial elliptic operators for nonsmooth data by using the Laplace transform method, where τ is the time step size. This convergence order is higher than the convergence orders of the popular convolution quadrature methods (e.g., Lubich’s convolution methods) and L-type methods (e.g., L1 method), which have only O(τ) convergence for the nonsmooth data. Numerical examples are given to verify the robustness of the time discretization schemes with respect to data regularity.
    • Error Estimates of a Continuous Galerkin Time Stepping Method for Subdiffusion Problem

      Yan, Yuyuan; Egwu, Bernard A.; Liang, Zongqi; Yan, Yubin; orcid: 0000-0002-5686-5017; email: y.yan@chester.ac.uk (Springer US, 2021-07-29)
      Abstract: A continuous Galerkin time stepping method is introduced and analyzed for subdiffusion problem in an abstract setting. The approximate solution will be sought as a continuous piecewise linear function in time t and the test space is based on the discontinuous piecewise constant functions. We prove that the proposed time stepping method has the convergence order O(τ1+α), α∈(0, 1) for general sectorial elliptic operators for nonsmooth data by using the Laplace transform method, where τ is the time step size. This convergence order is higher than the convergence orders of the popular convolution quadrature methods (e.g., Lubich’s convolution methods) and L-type methods (e.g., L1 method), which have only O(τ) convergence for the nonsmooth data. Numerical examples are given to verify the robustness of the time discretization schemes with respect to data regularity.