Heart Disease Detection using Vision-Based Transformer Models from ECG Images

Heart disease, also known as cardiovascular disease, is a prevalent and critical medical condition characterized by the impairment of the heart and blood vessels, leading to various complications such as coronary artery disease, heart failure, and myocardial infarction. The timely and accurate detection of heart disease is of paramount importance in clinical practice. Early identification of individuals at risk enables proactive interventions, preventive measures, and personalized treatment strategies to mitigate the progression of the disease and reduce adverse outcomes. In recent years, the field of heart disease detection has witnessed notable advancements due to the integration of sophisticated technologies and computational approaches. These include machine learning algorithms, data mining techniques, and predictive modeling frameworks that leverage vast amounts of clinical and physiological data to improve diagnostic accuracy and risk stratification. In this work, we propose to detect heart disease from ECG images using cutting-edge technologies, namely vision transformer models. These models are Google-Vit, Microsoft-Beit, and Swin-Tiny. To the best of our knowledge, this is the initial endeavor concentrating on the detection of heart diseases through image-based ECG data by employing cuttingedge technologies namely, transformer models. To demonstrate the contribution of the proposed framework, the performance of vision transformer models are compared with state-of-the-art studies. Experiment results show that the proposed framework exhibits remarkable classification results.

A Propagation-model Empowered Solution for Blind-Calibration of Sensors

Calibration of sensors is a major challenge especially in inexpensive sensors and sensors installed in inaccessible locations. The feasibility of calibrating sensors without the need for a standard sensor is called blind calibration. There is very little work in the open literature on totally blind calibration. In this work we model the sensing process as a combination of two processes, viz. propagation of the event through the environment to the sensor and measurement process in the sensor. Based on this, we propose a unique method for calibration in two flavours, viz semi-blind and completely-blind calibration. We show limited results based on simulation showing encouraging results.

Analysis of Arctic Buoy Dynamics using the Discrete Fourier Transform and Principal Component Analysis

Sea-Ice drift affects various global processes including the air-sea-ice energy system, numerical ocean modelling, and maritime activity in the polar regions. Drift has been investigated via various technologies ranging from satellite based systems to ship or ice-borne processes. This paper analyses the dynamics of sea-drift in the Arctic over 2019-2021 by Fourier Analysis and Principal Component Analysis of displacement data generated from the drift tracks of Ice-Tethered Profilers. We show that the frequency characteristics of drift support the notion that it is a function of both slowly varying processes, and higher frequency, random, forcing. In addition, we show that displacement data features high correlation between deployment locations and, consequently, suggest that there is scope for the optimisation of profiler deployment locations and for the reduction in number of instruments required to capture the displacement characteristics of drift.

Sensing the Environment with 5G Scattered Signals (5G-CommSense): A Feasibility Analysis

By making use of the sensors and AI (SensAI) algorithms for a specialized task, Application Specific INstrumentation (ASIN) framework uses less computational overhead and gives a good performance. This work evaluates the feasibility of the ASIN framework dependent Communication based Sensing (CommSense) system using 5th Generation New Radio (5G NR) infrastructure. Since our proposed system is backed up by 5G NR infra, this system is termed as 5G-CommSense. In this paper, we have used NR channel models specified by the 3rd Generation Partnership Project (3GPP) and added white Gaussian noise (AWGN) to vary the signal to noise ratio at the receiver. Finally, from our simulation result, we conclude that the proposed system is practically feasible.

Review of Methods for Handling Class-Imbalanced in Classification Problems

Learning classifiers using skewed or imbalanced datasets can occasionally lead to classification issues; this is a serious issue. In some cases, one class contains the majority of examples while the other, which is frequently the more important class, is nevertheless represented by a smaller proportion of examples. Using this kind of data could make many carefully designed machine-learning systems ineffective. High training fidelity was a term used to describe biases vs. all other instances of the class. The best approach to all possible remedies to this issue is typically to gain from the minority class. The article examines the most widely used methods for addressing the problem of learning with a class imbalance, including data-level, algorithm-level, hybrid, cost-sensitive learning, and deep learning, etc. including their advantages and limitations. The efficiency and performance of the classifier are assessed using a myriad of evaluation metrics.

Machine Learning based Extraction of Boundary Conditions from Doppler Echo Images for Patient Specific Coarctation of the Aorta: Computational Fluid Dynamics Study

Purpose- Coarctation of the Aorta (CoA) patient-specific computational fluid dynamics (CFD) studies in resource constrained settings are limited by the available imaging modalities for geometry and velocity data acquisition. Doppler echocardiography has been seen as a suitable velocity acquisition modality due to its higher availability and safety. This study aimed to investigate the application of classical machine learning (ML) methods to create an adequate and robust approach for obtaining boundary conditions (BCs) from Doppler Echocardiography images, for haemodynamic modeling using CFD. Methods- Our proposed approach combines ML and CFD to model haemodynamic flow within the region of interest. With the key feature of the approach being the use of ML models to calibrate the inlet and outlet boundary conditions (BCs) of the CFD model. The key input variable for the ML model was the patients heart rate as this was the parameter that varied in time across the measured vessels within the study. ANSYS Fluent was used for the CFD component of the study whilst the scikit-learn python library was used for the ML component. Results- We validated our approach against a real clinical case of severe CoA before intervention. The maximum coarctation velocity of our simulations were compared to the measured maximum coarctation velocity obtained from the patient whose geometry is used within the study. Of the 5 ML models used to obtain BCs the top model was within 5\% of the measured maximum coarctation velocity. Conclusion- The framework demonstrated that it was capable of taking variations of the patients heart rate between measurements into account. Thus, enabling the calculation of BCs that were physiologically realistic when the heart rate was scaled across each vessel whilst providing a reasonably accurate solution.

Sea Ice Concentration Estimation Techniques Using Machine Learning: An End-To-End Workflow for Estimating Concentration Maps from SAR Images

Sea ice concentration is an important metric used to characterize polar sea ice behavior. Understanding this behavior and accurately representing it is of critical importance for climate science research, and also has important uses in the context of maritime navigation. An end-to-end workflow for generating learned concentration estimation models from synthetic aperture radar data, trained on existing passive microwave data, is presented here. A novel objective function was introduced to account for uncertainty in the passive microwave measurements, which can be extended to account for arbitrary sources of error in the training data, and a recent set of in-situ observations was used to evaluate the reliability of the chosen passive microwave concentration estimation model. Google Colaboratory was used as the development platform, and all notebooks, training data, and trained models are available on GitHub. This chapter is an overview of the most interesting aspects of this investigation, and a detailed report is also available on GitHub.

Supervised Neural Networks for RFI Flagging

Neural network (NN) based methods are applied to the detection of radio frequency interference (RFI) in post-correlation,post-calibration time/frequency data. While calibration doesaffect RFI for the sake of this work a reduced dataset inpost-calibration is used. Two machine learning approachesfor flagging real measurement data are demonstrated usingthe existing RFI flagging technique AOFlagger as a groundtruth. It is shown that a single layer fully connects networkcan be trained using each time/frequency sample individuallywith the magnitude and phase of each polarization and Stokesvisibilities as features. This method was able to predict aBoolean flag map for each baseline to a high degree of accuracy achieving a Recall of 0.69 and Precision of 0.83 and anF1-Score of 0.75.

Machine Learning Techniques to Detect and Characterise Whistler Radio Waves

Lightning strokes create powerful electromagnetic pulses that routinely cause very low frequency (VLF) waves to propagate across hemispheres along geomagnetic field lines. VLF antenna receivers can be used to detect these whistler waves generated by these lightning strokes. The particular time/frequency dependence of the received whistler wave enables the estimation of electron density in the plasmasphere region of the magnetosphere. Therefore the identification and characterisation of whistlers are important tasks to monitor the plasmasphere in real-time and to build large databases of events to be used for statistical studies. The current state of the art in detecting whistler is the Automatic Whistler Detection (AWD) method developed by Lichtenberger (2009). This method is based on image correlation in 2 dimensions and requires significant computing hardware situated at the VLF receiver antennas (e.g. in Antarctica). The aim of this work is to develop a machine learning-based model capable of automatically detecting whistlers in the data provided by the VLF receivers. The approach is to use a combination of image classification and localisation on the spectrogram data generated by the VLF receivers to identify and localise each whistler. The data at hand has around 2300 events identified by AWD at SANAE and Marion and will be used as training, validation, and testing data. Three detector designs have been proposed. The first one using a similar method to AWD, the second using image classification on regions of interest extracted from a spectrogram, and the last one using YOLO, the current state of the art in object detection. It has been shown that these detectors can achieve a misdetection and false alarm of less than 15% on Marion's dataset.

WEBCA: Weakly-Electric-Fish Bioinspired Cognitive Architecture

Neuroethology has been an active field of study for more than a century now. Out of some of the most interesting species that has been studied so far, weakly electric fish is a fascinating one. It performs communication, echo-location and inter-species detection efficiently with an interesting configuration of sensors, neu-rons and a simple brain. In this paper we propose a cognitive architecture inspired by the way these fishes handle and process information. We believe that it is eas-ier to understand and mimic the neural architectures of a simpler species than that of human. Hence, the proposed architecture is expected to both help research in cognitive robotics and also help understand more complicated brains like that of human beings.