Polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs) prediction model based on limited peat samples using an evolved artificial neural network.

Polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs) are involuntary by-products of incomplete combustion and are highly toxic to humans and the environment. The Malaysian peat is often acidic or extremely acidic having high levels of chlorine and/or other organic acids that act as catalysts or precursors in PCDD/Fs formation. This study aims to predict PCDD/Fs emissions in peat soil using an artificial neural network (ANN) approach based on limited emission data and selected physico-chemical properties. The ANN's prediction performance is affected by uncertainties in its initial connection weights. To improve prediction performance, an optimisation algorithm, termed differential evolution (DE), is used to optimise the ANN's initial connection weights and bias. The study adopts several ANNs with fixed architecture to predict PCDD/Fs emissions, each consisting of a multilayer perceptron (MLP) with a backpropagation algorithm. Eight input variables and one output variable were adopted to train and test various neural network architectures using real-world datasets. The model optimisation procedure was conducted to ascertain the network architecture with the best predictive accuracy. The evolved ANN based on 5 hidden neurons, with the assistance of self-adaptive ensemble-based differential evolution with enhanced population sizing (SAEDE-EP), successfully produced the lowest MSEtest (6.1790 × 10-3) and highest R2 (0.97447) based on the mean among the other HNs. An evolutionary-optimised ANN-based methodology is a viable solution to predict PCDD/Fs in peat soil. It is cost-effective for pollution control, environmental monitoring and capable of aiding authorities prevent PCDD/Fs exposure, e.g., during a fire.

A variable dimensional chaotic map-based medical image encryption algorithm with multi-mode.

Since the COVID-19 pandemic, telemedicine or non-face-to-face medicine has increased significantly. In practice, various types of medical images are essential to achieve effective telemedicine. Medical image encryption algorithms play an irreplaceable role in the fast and secure transmission and storage of these medical images. However, most of the existing medical image encryption algorithms are full encryption algorithms, which are inefficient and time-consuming, so they are not suitable for emergency medical scenarios. To improve the efficiency of encryption, a small number of works have focused on partial or selective encryption algorithms for medical images, in which different levels of encryption strategies were adopted for different information content regions of medical images. However, these encryption algorithms have inadequate security more or less. In this paper, based on the Logistic map, we designed an improved variable dimension map. Then, an encryption algorithm for medical images was proposed based on it. This algorithm has two modes: (1) full encryption mode and (2) semi-full encryption mode, which can better adapt to different medical scenarios, respectively. In full encryption mode, all pixels of medical images are encrypted by using the confusion-diffusion structure. In semi-full encryption mode, the region of interest of medical images is extracted. The confusion was first adopted to encrypt the region of interest, and then, the diffusion was adopted to encrypt the entire image. In addition, no matter which encryption mode is used, the algorithm provides the function of medical image integrity verification. The proposed algorithm was simulated and analyzed to evaluate its effectiveness. The results show that in semi-full encryption mode, the algorithm has good security performance and lower time consumption; while in full encryption mode, the algorithm has better security performance and is acceptable in time.

Convolutional neural network optimized by differential evolution for electrocardiogram classification.

The Coronavirus disease 2019, or COVID-19, has shifted the medical paradigm from face-to-face to telehealth. Telehealth has become a vital resource to contain the virus spread and ensure the continued care of patients. In terms of preventing cardiovascular diseases, automating electrocardiogram (ECG) classification is a promising telehealth intervention. The healthcare service ensures that patient care is appropriate, comfortable, and accessible. Convolutional neural networks (CNNs) have demonstrated promising results in ECG categorization, which require high accuracy and short training time to ensure healthcare quality. This study proposes a one-dimensional-CNN (1D-CNN) arrhythmia classification based on the differential evolution (DE) algorithm to optimize the accuracy of ECG classification and training time. The performance of 1D-CNNs of different activation functions are optimized based on the standard DE algorithm. Finally, based on MIT-BIH and SCDH arrhythmia databases, the performances of optimized and unoptimized 1D-CNN are compared and analysed. Results show that the 1D-CNN optimized by the DE has higher accuracy in heartbeats classification. The optimized 1D-CNN improves from 97.6% to 99.5% on MIT-BIH and from 80.2% to 88.5% on SCDH. Therefore, the optimized 1D-CNN shows improvements of 1.9% and 8.3% in the two datasets, respectively. In addition, compared with the unoptimized 1D-CNN based on the same parameter settings, the optimized 1D-CNN has less training time. Under the conditions of ReLU function and 10 epochs, the training takes 9.22 s on MIT-BIH and 10.35 s on SCDH, reducing training time by 67.2% and 64.2%, respectively.

A plain-image correlative semi-selective medical image encryption algorithm using enhanced 2D-logistic map.

Modern medical examinations have produced a large number of medical images. It is a great challenge to transmit and store them quickly and securely. Existing solutions mainly use medical image encryption algorithms, but these encryption algorithms, which were developed for ordinary images, are time-consuming and must cope with insufficient security considerations when encrypting medical images. Compared with ordinary images, medical images can be divided into the region of interest and the region of background. In this paper, based on this characteristic, a plain-image correlative semi-selective medical image encryption algorithm using the enhanced two dimensional Logistic map was proposed. First, the region of interest of a plain medical image is permuted at the pixel level, then for the whole medical image, substitution is performed pixel by pixel. An ideal compromise between encryption speed and security can be achieved by full-encrypting the region of interest and semi-encrypting the region of background. Several main types of medical images and some normal images were selected as the samples for simulation, and main image cryptanalysis methods were used to analyze the results. The results showed that the cipher-images have a good visual quality, high information entropy, low correlation between adjacent pixels, as well as uniformly distribute histogram. The algorithm is sensitive to the initial key and plain-image, and has a large keyspace and low time complexity. The time complexity is lower when compared with the current medical image full encryption algorithm, and the security performance is better when compared with the current medical image selective encryption algorithm.

OES-Fed: a federated learning framework in vehicular network based on noise data filtering.

The Internet of Vehicles (IoV) is an interactive network providing intelligent traffic management, intelligent dynamic information service, and intelligent vehicle control to running vehicles. One of the main problems in the IoV is the reluctance of vehicles to share local data resulting in the cloud server not being able to acquire a sufficient amount of data to build accurate machine learning (ML) models. In addition, communication efficiency and ML model accuracy in the IoV are affected by noise data caused by violent shaking and obscuration of in-vehicle cameras. Therefore we propose a new Outlier Detection and Exponential Smoothing federated learning (OES-Fed) framework to overcome these problems. More specifically, we filter the noise data of the local ML model in the IoV from the current perspective and historical perspective. The noise data filtering is implemented by combining data outlier, K-means, Kalman filter and exponential smoothing algorithms. The experimental results of the three datasets show that the OES-Fed framework proposed in this article achieved higher accuracy, lower loss, and better area under the curve (AUC). The OES-Fed framework we propose can better filter noise data, providing an important domain reference for starting field of federated learning in the IoV.

Interactive Blood Vessel Segmentation from Retinal Fundus Image Based on Canny Edge Detector.

Optometrists, ophthalmologists, orthoptists, and other trained medical professionals use fundus photography to monitor the progression of certain eye conditions or diseases. Segmentation of the vessel tree is an essential process of retinal analysis. In this paper, an interactive blood vessel segmentation from retinal fundus image based on Canny edge detection is proposed. Semi-automated segmentation of specific vessels can be done by simply moving the cursor across a particular vessel. The pre-processing stage includes the green color channel extraction, applying Contrast Limited Adaptive Histogram Equalization (CLAHE), and retinal outline removal. After that, the edge detection techniques, which are based on the Canny algorithm, will be applied. The vessels will be selected interactively on the developed graphical user interface (GUI). The program will draw out the vessel edges. After that, those vessel edges will be segmented to bring focus on its details or detect the abnormal vessel. This proposed approach is useful because different edge detection parameter settings can be applied to the same image to highlight particular vessels for analysis or presentation.

A bibliometric of publication trends in medical image segmentation: Quantitative and qualitative analysis.

PURPOSE: Medical images are important in diagnosing disease and treatment planning. Computer algorithms that describe anatomical structures that highlight regions of interest and remove unnecessary information are collectively known as medical image segmentation algorithms. The quality of these algorithms will directly affect the performance of the following processing steps. There are many studies about the algorithms of medical image segmentation and their applications, but none involved a bibliometric of medical image segmentation. METHODS: This bibliometric work investigated the academic publication trends in medical image segmentation technology. These data were collected from the Web of Science (WoS) Core Collection and the Scopus. In the quantitative analysis stage, important visual maps were produced to show publication trends from five different perspectives including annual publications, countries, top authors, publication sources, and keywords. In the qualitative analysis stage, the frequently used methods and research trends in the medical image segmentation field were analyzed from 49 publications with the top annual citation rates. RESULTS: The analysis results showed that the number of publications had increased rapidly by year. The top related countries include the Chinese mainland, the United States, and India. Most of these publications were conference papers, besides there are also some top journals. The research hotspot in this field was deep learning-based medical image segmentation algorithms based on keyword analysis. These publications were divided into three categories: reviews, segmentation algorithm publications, and other relevant publications. Among these three categories, segmentation algorithm publications occupied the vast majority, and deep learning neural network-based algorithm was the research hotspots and frontiers. CONCLUSIONS: Through this bibliometric research work, the research hotspot in the medical image segmentation field is uncovered and can point to future research in the field. It can be expected that more researchers will focus their work on deep learning neural network-based medical image segmentation.