PyDTS: A Python Toolkit for Deep Learning Time Series Modelling.

In this article, the topic of time series modelling is discussed. It highlights the criticality of analysing and forecasting time series data across various sectors, identifying five primary application areas: denoising, forecasting, nonlinear transient modelling, anomaly detection, and degradation modelling. It further outlines the mathematical frameworks employed in a time series modelling task, categorizing them into statistical, linear algebra, and machine- or deep-learning-based approaches, with each category serving distinct dimensions and complexities of time series problems. Additionally, the article reviews the extensive literature on time series modelling, covering statistical processes, state space representations, and machine and deep learning applications in various fields. The unique contribution of this work lies in its presentation of a Python-based toolkit for time series modelling (PyDTS) that integrates popular methodologies and offers practical examples and benchmarking across diverse datasets.

HyperVein: A Hyperspectral Image Dataset for Human Vein Detection.

HyperSpectral Imaging (HSI) plays a pivotal role in various fields, including medical diagnostics, where precise human vein detection is crucial. HyperSpectral (HS) image data are very large and can cause computational complexities. Dimensionality reduction techniques are often employed to streamline HS image data processing. This paper presents a HS image dataset encompassing left- and right-hand images captured from 100 subjects with varying skin tones. The dataset was annotated using anatomical data to represent vein and non-vein areas within the images. This dataset is utilised to explore the effectiveness of dimensionality reduction techniques, namely: Principal Component Analysis (PCA), Folded PCA (FPCA), and Ward's Linkage Strategy using Mutual Information (WaLuMI) for vein detection. To generate experimental results, the HS image dataset was divided into train and test datasets. Optimum performing parameters for each of the dimensionality reduction techniques in conjunction with the Support Vector Machine (SVM) binary classification were determined using the Training dataset. The performance of the three dimensionality reduction-based vein detection methods was then assessed and compared using the test image dataset. Results show that the FPCA-based method outperforms the other two methods in terms of accuracy. For visualization purposes, the classification prediction image for each technique is post-processed using morphological operators, and results show the significant potential of HS imaging in vein detection.

Source Camera Identification Techniques: A Survey.

The successful investigation and prosecution of significant crimes, including child pornography, insurance fraud, movie piracy, traffic monitoring, and scientific fraud, hinge largely on the availability of solid evidence to establish the case beyond any reasonable doubt. When dealing with digital images/videos as evidence in such investigations, there is a critical need to conclusively prove the source camera/device of the questioned image. Extensive research has been conducted in the past decade to address this requirement, resulting in various methods categorized into brand, model, or individual image source camera identification techniques. This paper presents a survey of all those existing methods found in the literature. It thoroughly examines the efficacy of these existing techniques for identifying the source camera of images, utilizing both intrinsic hardware artifacts such as sensor pattern noise and lens optical distortion, and software artifacts like color filter array and auto white balancing. The investigation aims to discern the strengths and weaknesses of these techniques. The paper provides publicly available benchmark image datasets and assessment criteria used to measure the performance of those different methods, facilitating a comprehensive comparison of existing approaches. In conclusion, the paper outlines directions for future research in the field of source camera identification.

Microbiological Quality Estimation of Meat Using Deep CNNs on Embedded Hardware Systems.

Spectroscopic sensor imaging of food samples meta-processed by deep machine learning models can be used to assess the quality of the sample. This article presents an architecture for estimating microbial populations in meat samples using multispectral imaging and deep convolutional neural networks. The deep learning models operate on embedded platforms and not offline on a separate computer or a cloud server. Different storage conditions of the meat samples were used, and various deep learning models and embedded platforms were evaluated. In addition, the hardware boards were evaluated in terms of latency, throughput, efficiency and value on different data pre-processing and imaging-type setups. The experimental results showed the advantage of the XavierNX platform in terms of latency and throughput and the advantage of Nano and RP4 in terms of efficiency and value, respectively.

Contactless Camera-Based Heart Rate and Respiratory Rate Monitoring Using AI on Hardware.

Detecting vital signs by using a contactless camera-based approach can provide several advantages over traditional clinical methods, such as lower financial costs, reduced visit times, increased comfort, and enhanced safety for healthcare professionals. Specifically, Eulerian Video Magnification (EVM) or Remote Photoplethysmography (rPPG) methods can be utilised to remotely estimate heart rate and respiratory rate biomarkers. In this paper two contactless camera-based health monitoring architectures are developed using EVM and rPPG, respectively; to this end, two different CNNs, (Mediapipe's BlazeFace and FaceMesh) are used to extract suitable regions of interest from incoming video frames. These two methods are implemented and deployed on four off-the-shelf edge devices as well as on a PC and evaluated in terms of latency (in each stage of the application's pipeline), throughput (FPS), power consumption (Watt), efficiency (throughput/Watt), and value (throughput/cost). This work provides important insights about the computational costs and bottlenecks of each method on each hardware platform, as well as which platform to use depending on the target metric. One of our insights shows that the Jetson Xavier NX platform is the best platform in terms of throughput and efficiency, while Raspberry Pi 4 8 GB is the best platform in terms of value.

2D Transformations of Energy Signals for Energy Disaggregation.

The aim of Non-Intrusive Load Monitoring is to estimate the energy consumption of individual electrical appliances by disaggregating the overall power consumption that has been sampled from a smart meter at a house or commercial/industrial building. Last decade's developments in deep learning and the utilization of Convolutional Neural Networks have improved disaggregation accuracy significantly, especially when utilizing two-dimensional signal representations. However, converting time series' to two-dimensional representations is still an open challenge, and it is not clear how it influences the performance of the energy disaggregation. Therefore, in this article, six different two-dimensional representation techniques are compared in terms of performance, runtime, influence on sampling frequency, and robustness towards Gaussian white noise. The evaluation results show an advantage of two-dimensional imaging techniques over univariate and multivariate features. In detail, the evaluation results show that: first, the active and reactive power-based signatures double Fourier based signatures, as well as outperforming most of the other approaches for low levels of noise. Second, while current and voltage signatures are outperformed at low levels of noise, they perform best under high noise conditions and show the smallest decrease in performance with increasing noise levels. Third, the effect of the sampling frequency on the energy disaggregation performance for time series imaging is most prominent up to 1.2 kHz, while, above 1.2 kHz, no significant improvements in terms of performance could be observed.

Energy Disaggregation Using Elastic Matching Algorithms.

In this article an energy disaggregation architecture using elastic matching algorithms is presented. The architecture uses a database of reference energy consumption signatures and compares them with incoming energy consumption frames using template matching. In contrast to machine learning-based approaches which require significant amount of data to train a model, elastic matching-based approaches do not have a model training process but perform recognition using template matching. Five different elastic matching algorithms were evaluated across different datasets and the experimental results showed that the minimum variance matching algorithm outperforms all other evaluated matching algorithms. The best performing minimum variance matching algorithm improved the energy disaggregation accuracy by 2.7% when compared to the baseline dynamic time warping algorithm.

Update in COVID-19 in the intensive care unit from the 2020 HELLENIC Athens International symposium.

The 2020 International Web Scientific Event in COVID-19 pandemic in critically ill patients aimed at updating the information and knowledge on the COVID-19 pandemic in the intensive care unit. Experts reviewed the latest literature relating to the COVID-19 pandemic in critically ill patients, such as epidemiology, pathophysiology, phenotypes of infection, COVID-19 as a systematic infection, molecular diagnosis, mechanical ventilation, thromboprophylaxis, COVID-19 associated co-infections, immunotherapy, plasma treatment, catheter-related bloodstream infections, artificial intelligence for COVID-19, and vaccination. Antiviral therapy and co-infections are out of the scope of this review. In this review, each of these issues is discussed with key messages regarding management and further research being presented after a brief review of available evidence.

Colour Constancy for Image of Non-Uniformly Lit Scenes.

Digital camera sensors are designed to record all incident light from a captured scene, but they are unable to distinguish between the colour of the light source and the true colour of objects. The resulting captured image exhibits a colour cast toward the colour of light source. This paper presents a colour constancy algorithm for images of scenes lit by non-uniform light sources. The proposed algorithm uses a histogram-based algorithm to determine the number of colour regions. It then applies the K-means++ algorithm on the input image, dividing the image into its segments. The proposed algorithm computes the Normalized Average Absolute Difference (NAAD) for each segment and uses it as a measure to determine if the segment has sufficient colour variations. The initial colour constancy adjustment factors for each segment with sufficient colour variation is calculated. The Colour Constancy Adjustment Weighting Factors (CCAWF) for each pixel of the image are determined by fusing the CCAWFs of the segments, weighted by their normalized Euclidian distance of the pixel from the center of the segments. Results show that the proposed method outperforms the statistical techniques and its images exhibit significantly higher subjective quality to those of the learning-based methods. In addition, the execution time of the proposed algorithm is comparable to statistical-based techniques and is much lower than those of the state-of-the-art learning-based methods.

Spike pattern recognition by supervised classification in low dimensional embedding space.

Epileptiform discharges in interictal electroencephalography (EEG) form the mainstay of epilepsy diagnosis and localization of seizure onset. Visual analysis is rater-dependent and time consuming, especially for long-term recordings, while computerized methods can provide efficiency in reviewing long EEG recordings. This paper presents a machine learning approach for automated detection of epileptiform discharges (spikes). The proposed method first detects spike patterns by calculating similarity to a coarse shape model of a spike waveform and then refines the results by identifying subtle differences between actual spikes and false detections. Pattern classification is performed using support vector machines in a low dimensional space on which the original waveforms are embedded by locality preserving projections. The automatic detection results are compared to experts' manual annotations (101 spikes) on a whole-night sleep EEG recording. The high sensitivity (97 %) and the low false positive rate (0.1 min-1), calculated by intra-patient cross-validation, highlight the potential of the method for automated interictal EEG assessment.

Real-Time Management of Multimodal Streaming Data for Monitoring of Epileptic Patients.

New generation of healthcare is represented by wearable health monitoring systems, which provide real-time monitoring of patient's physiological parameters. It is expected that continuous ambulatory monitoring of vital signals will improve treatment of patients and enable proactive personal health management. In this paper, we present the implementation of a multimodal real-time system for epilepsy management. The proposed methodology is based on a data streaming architecture and efficient management of a big flow of physiological parameters. The performance of this architecture is examined for varying spatial resolution of the recorded data.

Joint spatial-spectral feature space clustering for speech activity detection from ECoG signals.

Brain-machine interfaces for speech restoration have been extensively studied for more than two decades. The success of such a system will depend in part on selecting the best brain recording sites and signal features corresponding to speech production. The purpose of this study was to detect speech activity automatically from electrocorticographic signals based on joint spatial-frequency clustering of the ECoG feature space. For this study, the ECoG signals were recorded while a subject performed two different syllable repetition tasks. We found that the optimal frequency resolution to detect speech activity from ECoG signals was 8 Hz, achieving 98.8% accuracy by employing support vector machines as a classifier. We also defined the cortical areas that held the most information about the discrimination of speech and nonspeech time intervals. Additionally, the results shed light on the distinct cortical areas associated with the two syllables repetition tasks and may contribute to the development of portable ECoG-based communication.