Search

Neck pain is one of the most common symptoms of cervical spine disease, and segmenting neck muscles to create volumetric measurements may assist clinical diagnosis. While image registration is used to segment medical images, registration is highly challenging for neck muscles due to their tight proximity, shape and size variations among subjects, and similar appearance. These challenges cause conventional multi resolution-based registration methods to be trapped in local minima due to their low degree of freedom geometrical transforms. This article presents a novel object-constrained hierarchical registration framework for aligning inter-subject neck muscles. First, to handle large scale local minima, the proposed framework uses a coarse registration technique, which optimizes the new edge position difference (EPD) similarity measure, to align large mismatches. Also, a new transformation based on the discrete periodic spline wavelet (DPSW), affine and free-form-deformation (FFD) transformations are exploited. Second, to avoid monotonous nature of using transformations in multiple stages, a fine registration technique is designed for aligning small mismatches. This technique uses a double-pushing system by changing edges in the EPD and switching transformation resolutions. The EPD helps in both coarse and fine techniques to implement object-constrained registration via controlling edges, which is not possible when using traditional similarity measures. Experiments are performed on clinical 3D magnetic resonance imaging (MRI) scans of the neck, with the results showing that the EPD is more effective than the mutual information (MI) and sum of squared difference (SSD) measure in terms of volumetric dice similarity coefficient (DSC). Additionally, the proposed method is compared with the diffeomorphic Demons and SyN state-of-the-art approaches with ablation studies in inter-subject deformable registration. The proposed method achieves better accuracy, robustness and consistency than the reference methods, with an average volumetric DSC of 0.7029 compared to 0.6654 and 0.6606 for the Demons and SyN methods, respectively.

The accuracy of current natural scene text recognition algorithms is limited by the poor performance of character recognition methods for these images. The complex backgrounds, variations in the writing, text size, orientations, low resolution and multi-language text make recognition of text in natural images a complex and challenging task. Conventional machine learning and deep learning-based methods have been developed that have achieved satisfactory results, but character recognition for cursive text such as Arabic and Urdu scripts in natural images is still an open research problem. The characters in the cursive text are connected and are difficult to segment for recognition. Variations in the shape of a character due to its different positions within a word make the recognition task more challenging than non-cursive text. Optical character recognition (OCR) techniques proposed for Arabic and Urdu scanned documents perform very poorly when applied to character recognition in natural images. In this paper, we propose a multi-scale feature aggregation (MSFA) and a multi-level feature fusion (MLFF) network architecture to recognize isolated Urdu characters in natural images. The network first aggregates multi-scale features of the convolutional layers by up-sampling and addition operations and then combines them with the high-level features. Finally, the outputs of the MSFA and MLFF networks are fused together to create more robust and powerful features. A comprehensive dataset of segmented Urdu characters is developed for the evaluation of the proposed network models. Synthetic text on the patches of images with real natural scene backgrounds is generated to increase the samples of infrequently used characters. The proposed model is evaluated on the Chars74K and ICDAR03 datasets. To validate the proposed model on the new Urdu character image dataset, we compare its performance with the histogram of oriented gradients (HoG) method. The experimental results show that the aggregation of multi-scale and multilevel features and their fusion is more effective, and outperforms other methods on the Urdu character image and Chars74K datasets.

The Lightfield Single Photon Emission Computed Tomography (L-SPECT) system is developed to overcome some of the drawbacks in conventional SPECT by applying the idea of plenoptic imaging. This system displayed improved performance in terms of reduced information loss and scanning time when compared to the SPECT system which has a conventional collimator. The SPECT system is transformed into L-SPECT system by replacing the conventional collimators with micro-range multi-pinhole arrays. The field of view (FOV) of the L-SPECT system is enhanced by reshaping the detector head into ring-type by tiling small detector modules. The L-SPECT system with modular partial ring detectors (MPRD L-SPECT) exhibits cylindrical artefacts during volumetric reconstruction. Hence, here the work is focused to remove the cylindrical artefacts in the reconstruction of the examined objects by changing the scanning orbit. The enhancement is done such that the center of scanning of the L-SPECT system with MPRD L-SPECT is shifted at different values. The reconstruction quality of MPRD L-SPECT with and without center shifting is evaluated in terms of the Full Width at Half Maximum (FWHM) and Modulation Transfer Function (MTF). Moreover, the visual comparison is also examined. The results indicate that center shifting of MPRD L-SPECT overcomes the problem of the central artefact with improved FWHM. By increasing MPRD's scanning center shift gap, the spatial resolution can be further improved.

We describe a novel approach to regulatory compliance decision support that leverages an asset-based cyber security risk management approach following ISO 27005, and illustrate this with an example based on the General Data Protection Regulation (GDPR). Previous work in regulatory compliance modelling and decision support utilises semantic vocabularies and reasoning techniques, and our approach has the primary benefit that regulatory compliance is integrated with other domains of risk management, so that insight from domains such as cyber security combined with regulatory compliance for privacy can be gained based on a single model of the user's sociotechnical infrastructure. The main contributions of this paper are: to show how regulatory requirements may be modelled as “compliance threats” in an asset-based risk management framework; to illustrate mapping from the GDPR's legal text to domain assets, processes and relationships, compliance threats and control strategies to mitigate the threats; to show how the threats are triggered via recognition patterns based on asset configurations; to illustrate how the different types of regulatory requirement, e.g. obligations, prohibitions and derogating conditions, are represented in such a scheme; and finally to describe how to model causal dependencies between choices made to address a compliance threat and downstream additional compliance requirements.

Immersive media representation format based on point clouds has underpinned significant opportunities for extended reality applications. Point cloud in its uncompressed format require very high data rate for storage and transmission. The video based point cloud compression (V-PCC) technique projects a dynamic point cloud into geometry and texture video sequences. The projected texture video is then coded using modern video coding standard like HEVC. Since the properties of projected texture video frames are different from traditional video frames, HEVC-based commonality modeling can be inefficient. An improved commonality modeling technique is proposed that employs edge position difference oriented motion model. Experimental results show that the proposed commonality modeling technique can yield savings in bit rate of up to 3.15% over the V-PCC HEVC reference encoder.

Ahstract- The ability to detect and analyze changes or understand the scene while navigating close to buildings is very important for autonomous aerial and ground vehicle based surveillance applications. For this, the latest textured 3D scan of the platform's view frustum has to be placed accurately in the context of a big map like a Structure from Motion (SfM) map of the region. However, due to the drift in the camera trajectory, the scans are usually not aligned with the SfM model. This paper proposes a novel registration algorithm that aligns the 3D scan using known 2D images of the SfM model. The proposed 3D-2D registration method uses a heuristic approach which first performs a robust 2D-2D registration between the projection of the 3D scan and the SfM images and then calculates the 3D alignment parameters by combining registration results of multiple camera views of the SfM model. The results presented compare the robustness of the proposed registration techniques with traditional approaches.

The segmentation of neck muscles is useful for the diagnoses and planning of medical interventions for neck pain-related conditions such as whiplash and cervical dystonia. Neck muscles are tightly grouped, have similar appearance to each other and display large anatomical variability between subjects. They also exhibit low contrast with background organs in magnetic resonance (MR) images. These characteristics make the segmentation of neck muscles a challenging task. Due to the significant success of the U-Net architecture for deep learning-based segmentation, numerous versions of this approach have emerged for the task of medical image segmentation. This paper presents an evaluation of 10 U-Net CNN approaches, 6 direct (U-Net, CRF-Unet, A-Unet, MFP-Unet, R2Unet and U-Net++) and 4 modified (R2A-Unet, R2A-Unet++, PMS-Unet and MS-Unet). The modifications are inspired by recent multi-scale and multi-stream techniques for deep learning algorithms. T1 weighted axial MR images of the neck, at the distal end of the C3 vertebrae, from 45 subjects with real-time data augmentation were used in our evaluation of neck muscle segmentation approaches. The analysis of our numerical results indicates that the R2Unet architecture achieves the best accuracy.

During the spread of an infectious disease such as COVID-19, the identification of human factors that affect the spread is a really important area of research. These factors directly impact the spread of such a disease and are important in identifying the various regions that are at a higher risk than others. This allows for an optimal distribution of resources according to predicted demand. Traditional infectious modeling techniques are good at representing the spread and can incorporate multiple factors that resemble real-life scenarios. The primary issue here is the identification of relevant variables. In this study, a residual analysis is presented to downsize the dataset available and shortlist the variables classified as absolutely necessary for disease modeling. The performance of different datasets is evaluated using an Artificial Neural Network and regression analysis. The results show that the drop in performance using the reduced dataset is reasonable as it is very difficult to obtain a perfect dataset covering only necessary variables. This approach can be automated in the future as it offers a small dataset containing a few variables against a large dataset with possibly hundreds of variables.

Many machine learning methods are being developed to predict the spread of COVID - 19. This paper focuses on the expansion of inputs that may be considered in these models. A correlation matrix is used to identify those variables with the highest correlation to COVID - 19 cases. These variables are then used and compared in three methods that predict future cases: a Support Vector Machine Regression (SVR), Multidimensional Regression with Interactions, and the Stepwise Regression method. All three methods predict a rise in cases similar to the actual rise in cases, and importantly, are all able to predict to a certain degree the unexpected dip in cases on the 10th and 11th day of prediction.

In Tactical Combat Casualty Care (TCCC), medics perform Role 1 care for battlefield casualties at point of injury by stabilizing them and transporting them to field care facilities such as a Battalion Aid Station (Role 2) or Field Hospital (Role 3) where clinicians provide critical care. Care provider experience and ability vary, and training in the field can help to improve recall and performance of infrequently used critical care skills. This becomes more necessary during Prolonged Field Care (PFC) when evacuation is not immediately available and more complex treatment may be required. Our Trauma Triage Treatment and Training Decision Support (4TDS) project has developed a decision support system (DSS) for Roles 1 and 2. As an application on a Android smart phone and tablet, 4TDS includes training scenarios in skills such as shock identification and management. 4TDS pairs with various vital signs sensors that can stream data for a machine learning algorithm that can detect the probability of shock in a casualty. A "silent test" is comparing algorithm performance with actual clinical diagnoses at Mayo Clinic, Rochester, MN. Usability assessment in an austere field setting will enable us to determine medic and clinician acceptance of 4TDS and how well it supports their decision making. Faster, more accurate decisions can improve TCCC patient care under conditions in which delays can increase morbidity and mortality.

Efficient video compression relies heavily on mitigating the temporal redundancy that exists between successive video frames. This is achieved through effective motion modelling. In conventional video coding standards, the motion of the current frame is modelled from the neighbouring frames using block-based motion estimation techniques. However, as the motion discontinuities are tied to the moving objects in a video frame, the block-based techniques are unable to model the actual motion of individual objects. In this paper, an object-based hierarchical motion estimation and prediction technique for high-efficiency video coding (HEVC) is proposed. We use an edge position difference (EPD) similarity measure, which has the ability to align the largest object in the frames, to estimate the motion of the object in the current frame from the neighbouring one. In other words, it estimates the largest object’s motion instead of the whole frame’s global motion. The proposed method gradually models all of the objects’ motions and establishes a prediction of the current frame. The predicted frame is then exploited as an additional reference frame in the HEVC compression algorithm. Our experimental results demonstrate that our proposed approach achieves a bit rate savings with a peak signal to noise ratio (PSNR) gain over the HEVC standard.

Video data storage and transmission cost can be reduced by minimizing the temporally redundant information among frames using an appropriate motion-compensated prediction technique. In the current video coding standard, the neighbouring frames are exploited to predict the motion of the current frame using global motion estimation-based approaches. However, the global motion estimation of a frame may not produce the actual motion of individual objects in the frame as each of the objects in a frame usually has its own motion. In this paper, an edge-based motion estimation technique is presented that finds the motion of each object in the frame rather than finding the global motion of that frame. In the proposed method, edge position difference (EPD) similarity measure-based image registration between the two frames is applied to register each object in the frame. A superpixel search is then applied to segment the registered object. Finally, the proposed edge-based image registration technique and Demons algorithm are applied to predict the objects in the current frame. Our experimental analysis demonstrates that the proposed algorithm can estimate the motions of individual objects in the current frame accurately compared to the existing global motion estimation-based approaches.

Premature births are highly prevalent; world-wide, one in every ten births is premature. A common complication of premature birth is the inability of the newborns to regulate their own body temperature (hypothermia). Kangaroo Mother Care (KMC) is an effective strategy to mitigate this problem but has challenges. Namely, KMC requires constant engagement from the caregiver in order to keep the baby warm and to monitor the baby's temperature every 4 hours. In this report we describe NeoWarm, a device designed to address these challenges via continuous integrated warming and automated temperature monitoring of the infant, both during KMC and when the caregiver takes a break. Pilot studies demonstrate the ability of the device to warm a simulated newborn from moderate hypothermia to ideal core temperatures within two hours.

With the growing trend towards personalized health, wearable fitness trackers have become a staple of the consumer electronics industry. As the prevalence of such devices booms, the medical community has been compelled to investigate the potential of such devices and explore how they can be used to create positive clinical outcomes. In this report we detail a smart-ring capable of determining heart rate (HR), respiratory rate (RR), blood oxygen saturation (SpO 2) , and temperature. The ring implements a photoplethysmogram (PPG), electrocardiogram (ECG), and thermistor to attain these metrics. After evaluation, significant correlation was found between the experimentally reported HR and RR recordings with their respective standards (p 2 had trial-dependent similarities with its reference standards, and temperature measurements were within expected values of normal skin temperature.

Agriculture is facing a period of unprecedented change resulting from the requirement to feed c. 10 billion mouths by 2050, while operating in a challenging business environment with increasing sustainability and consumer expectations. To meet these challenges many are looking to industry 4.0 technology as a solution, specifically the combination of Artificial Intelligence (AI) and in-field robotics. Although there have been successful prototypes from a functional perspective and promising start-ups, the complex nature of the agriculture system of systems creates a significant risk of mass adoption failure. The main risk stems from the majority of the hundreds of products being designed in isolation using closed hardware and software systems that are all aiming to serve the one farmer or grower. The current silo’d technology adoption approach is heading towards problems with usability, availability, viability, interoperability and adaptability. This paper proposes a new time constrained System of Systems (SoS) discovery process and supporting SoS Canvas, developed with the objective of increasing the adoption of SoS principles with non-technical stakeholders. The time constrained SoS discovery process consists of raising awareness, capturing data, modelling the SoS, modelling the System of Interest (SoI) and providing recommendations, often based on common challenges and opportunities from the SoS Canvas. The outcome of this initial case study prototyping the process, is a set of SysML diagrams demonstrating how the constituent systems currently contribute to emergent properties and recommendations to improve both the SoI and the SoS. Future research will focus on interoperability within agricultural technology (Agritech) and the potential for a SoS level digital twin, providing the capability to manage and optimise emergent properties for the betterment of all.