Search

Evolutionary computation is derived from the simulation of natural selection and genetic processes in biological evolution. This approach provides a method for optimizing the structure and parameters of neural networks. When combined with neural networks, forming what's termed as evolutionary computation based neural networks, it offers a systematic approach to optimize neural network models in diverse applications. In this study, we introduce a method that employs differential evolution algorithms to optimize parameters of convolutional neural network (CNN) for music emotion recognition tasks. This method optimizes the initial weights of the CNN, aiming to achieve near-global optimal solutions and expedite network convergence. Comparative experiments indicate that the proposed approach effectively identifies optimal parameters and structures for CNN, suggesting potential advancements in automated music emotion recognition. • This paper improves differential evolution algorithm via combining five key operations. • This paper employs differential evolution algorithm to optimize parameters of CNNs for music emotion recognition tasks. • The proposed DE-CNN approach effectively identifies optimal parameters and structures for CNNs. [ABSTRACT FROM AUTHOR]

Determining the correct mushroom species with the necessary ecological characteristics is critical to continue mushroom production, which is essential in gastronomy. The mushroom farmers and collectors technique may help identify toxic mushrooms by detecting poisonous mushrooms using images of different mushroom species with distinctive morphological features. However, it can be not easy to distinguish between species. This paper used a dataset of 6714 mushroom images obtained from nine different mushroom species to classify the mushroom species. For a more straightforward comprehension of mushroom images and feature extraction by reanalysis of data sets, data visualization was performed using Grad-CAM, LIME, and Heatmap methods. Residual block-based Convolutional Neural Network (CNN) architectures are trained to automatically classify the concatenated feature map obtained from the Grad-CAM, LIME, and Heatmap methods. After extracting the deep features of the images from each architecture, the Atom Search Optimization (ASO) algorithm has been used to select the most distinctive features. The 6714×9000 size of the concatenated feature map was reduced to 6714×600 using the ASO algorithm. Classification results were evaluated using six different classifiers based on the feature map obtained to determine the mushroom species. The nine classes of mushroom species were classified successfully with 95.45 % accuracy using the proposed model with the ASO algorithm and KNN classifier. The methodology introduces novel visualization techniques for interpreting CNN-based models' decisions in mushroom species classification tasks. Using metaheuristics-based CNN models with multi-feature fusion techniques allows the model to leverage diverse sources of information, potentially enhancing its ability to discriminate between mushroom species and achieve higher classification accuracy than existing methods. This study can advance the mushroom species classification field by introducing new methodologies, improving classification accuracy, providing insights into model interpretability, and facilitating knowledge transfer to related fields. • The proposed model automatically detects mushroom species. • Grad-CAM, LIME, and Heatmap methods based on CNN architectures are used to visualize mushroom images. • The ASO algorithm has been successfully applied to residual block-based CNN models for feature selection. • The hybrid model classification accuracy for different species of mushrooms is 95.45 %. [ABSTRACT FROM AUTHOR]

In the past few years, convolutional neural networks (CNN) and Transformer have achieved great success in medical image segmentation, but they each have inevitable drawbacks. Among them, convolution operation is difficult to calculate the relationship between elements at a certain position and elements far away from the position in the feature map. However, Transformer tends to ignore the importance of local information when exploring the correlation between overall elements. In order to allow the network to acquire both the ability to explore local details and compute the correlation between distant elements, this paper proposes TransUNet++ based on TransUNet. Specifically, this paper proposes two modules, Big Model and Small Model, to explore the element relationship between feature maps. Among them, on the basis of the whole feature map as the basic unit, the Big model can not only calculate the correlation between distant elements in the feature map but also extract the detailed information of the local feature map. On the basis of taking 1/4 of the feature map as the basic unit, the Small model not only explores the correlation between distant elements but also extracts the local details of the feature map. We demonstrate on the Synapse multi-organ segmentation dataset(Synapse) and Automated cardiac diagnosis challenge dataset (ACDC) that using either the Big Model or the Small Model alone can improve the experimental results, and using the Big Model and the Small Model in parallel can achieve more optimal experimental results. Among them, in Synapse dataset, we achieved 80.87% dice score and 24.79% HD score, and in ACDC dataset, we achieved 91.41% dice score and 1.08% HD score. [Display omitted] • The Big Model can be used for remote modeling and local information extraction. • The Small models can be used for mid-range modeling and local detail extraction. • The advantages and disadvantages of the convolution kernel and Transformer by talked. [ABSTRACT FROM AUTHOR]

Aiming at the limitation of Hamiltonian neural network in high-dimensional data processing with non-observable physical quantity system, an embedded Hamiltonian dynamic evolutionary neural network model (EHDN) is proposed in this paper. First, the dynamic physical quantities of high dimensional nonlinear systems are represented by constructing Hamiltonian functions. Then a symplectic integrator is designed to realize the time-varying evolution of network features based on Hamiltonian regular equation and variational numerical integration. Finally, the Hamiltonian dynamic evolutionary neural network is embedded in the convolutional neural network (CNN) to realize the Hamiltonian simulation evolution of convolutional features. Experimental results on two different datasets, CIFAR and Fashion-Mnist, show that compared with existing state-of-the-art (SOTA) methods, EHDN model can combine the advantages of Hamiltonian dynamics and convolutional network to improve the recognition accuracy and training stability in high-dimensional image classification tasks. • An embeddable Hamiltonian neural network model is proposed to solve the degradation problem of very deep networks. • The high-dimensional image features are self-evolved by the latent Hamiltonian to reduce the hyperparametric constraints. • EHDN network architecture is flexible, and two evolution modes can be designed on demand. [ABSTRACT FROM AUTHOR]

The construction of smart grid is inseparable from the appropriate deployment of utility poles. China has many poles; if they are fully utilized, can they bring more power to the smart grid? Fortunately, thanks to advances in technologies such as artificial intelligence and the Internet of Things, can we use Google Earth to recognize utility poles and optimize the deployment of power facilities? Fundamentally, Google Earth generates static images. Faced with many static images, we need further image processing for high-definition recognition of utility poles. In addition to the current chaotic distribution of poles, we also need to optimize the deployment of power facilities. However, due to the resolution of Google Earth and the aerial photography angle, many backlight phenomena are not conducive to the recognition of utility poles. Therefore, this paper proposes a backlight image enhancement algorithm based on convolutional neural networks (CNN) and constructs a novel network architecture that integrates decomposition, restoration, and adjustment to recognize poles in high-definition under backlight. Furthermore, to solve the problems of the overflow of CNN parameters and unclear training effect, particle swarm optimization (PSO), the evolutionary computing-based machine learning (ECML) is used to search CNN parameters automatically and seek the optimal solution to achieve the optimization of the overall model. The experiments prove that the CNN-based image enhancement algorithm effectively recognizes the utility poles under different illumination and backlight. At the same time, the experimental results show that the PSO-based optimization method can optimize CNN parameters obviously, and the classification accuracy is increased. • An image enhancement algorithm based on CNN for Google Earth static images is proposed for recognition of utility poles. • A novel network architecture integrating decomposition, restoration, and adjustment is constructed. • A PSO-based method is proposed to automatically tune the parameters of CNN model. [ABSTRACT FROM AUTHOR]

Financial time series forecasting is undoubtedly the top choice of computational intelligence for finance researchers in both academia and the finance industry due to its broad implementation areas and substantial impact. Machine Learning (ML) researchers have created various models, and a vast number of studies have been published accordingly. As such, a significant number of surveys exist covering ML studies on financial time series forecasting. Lately, Deep Learning (DL) models have appeared within the field, with results that significantly outperform their traditional ML counterparts. Even though there is a growing interest in developing models for financial time series forecasting, there is a lack of review papers that solely focus on DL for finance. Hence, the motivation of this paper is to provide a comprehensive literature review of DL studies on financial time series forecasting implementation. We not only categorized the studies according to their intended forecasting implementation areas, such as index, forex, and commodity forecasting, but we also grouped them based on their DL model choices, such as Convolutional Neural Networks (CNNs), Deep Belief Networks (DBNs), and Long-Short Term Memory (LSTM). We also tried to envision the future of the field by highlighting its possible setbacks and opportunities for the benefit of interested researchers. • We reviewed all searchable articles of deep learning (DL) for financial time series forecasting. • RNN based DL models (LSTM and GRU included) are the most common. • We compared DL models according to their performances in different forecasted asset classes. • To best of our knowledge, this is the first comprehensive DL survey for financial time series forecasting. • We provided current status of DL in financial time series forecasting, also highlighted the future opportunities. [ABSTRACT FROM AUTHOR]

Through the commencement of the COVID-19 pandemic, the whole globe is in disarray and debating on unique approaches to stop this viral transmission. Masks are being worn by people all around the world as one of the preventative measures to avoid contracting this sickness. Although some people are following and adopting this precaution, others are not, despite official recommendations from the administration and public health organisations has been announced. In this paper DTLMV2 (Deep Transfer Learning MobileNetV2 for the objective of classification) is proposed - A face mask identification model that can reliably determine whether an individual is wearing a mask or not is suggested and implemented in this work. The model architecture employs the peruse of MobileNetV2, a lightweight Convolutional Neural Network (CNN) that requires less computing power and can be readily integrated into computer vision and mobile systems. The computer vision with MobileNet is required to formulate a low-cost mask detection system for a group of people in open spaces that can assist in determining whether a person is wearing a mask or not, as well as function as a surveillance system since it is effective on both real-time pictures and videos. The face recognition model obtained 97.01% accuracy on validation data, 98% accuracy on training data and 97.45% accuracy on testing data. [ABSTRACT FROM AUTHOR]

Ever since the outbreak of COVID-19, the entire world is grappling with panic over its rapid spread. Consequently, it is of utmost importance to detect its presence. Timely diagnostic testing leads to the quick identification, treatment and isolation of infected people. A number of deep learning classifiers have been proved to provide encouraging results with higher accuracy as compared to the conventional method of RT-PCR testing. Chest radiography, particularly for using X-ray images, is a prime imaging modality for detecting the suspected COVID-19 patients. However, the performance of these approaches still needs to be improved. In this paper, we propose a capsule network called COVID-WideNet for diagnosing COVID-19 cases using Chest X-ray (CXR) images. Experimental results have demonstrated that a discriminative trained, multi-layer capsule network achieves state-of-the-art performance on the COVIDx dataset. In particular, COVID-WideNet performs better than any other CNN based approaches for the diagnosis of COVID-19 infected patients. Further, the proposed COVID-WideNet has the number of trainable parameters that is 20 times less than that of other CNN based models. This results in a fast and efficient diagnosing COVID-19 symptoms, and with achieving the 0.95 of Area Under Curve (AUC), 91% of accuracy, sensitivity and specificity, respectively. This may also assist radiologists to detect COVID and its variant like delta. • Proposed COVID-WideNet a fast and efficient Capsule network for COVID-19 detection. • Effective in real-time COVID-19 detection due to 20 times less parameters from CNN. • Discriminative approach towards COVID-19 detection; eliminates limitations of CNN. • Result shows 91% sensitivity, specificity, accuracy and 0.95 of area under curve. [ABSTRACT FROM AUTHOR]

Coronavirus Disease 2019 (COVID-19) had already spread worldwide, and healthcare services have become limited in many countries. Efficient screening of hospitalized individuals is vital in the struggle toward COVID-19 through chest radiography, which is one of the important assessment strategies. This allows researchers to understand medical information in terms of chest X-ray (CXR) images and evaluate relevant irregularities, which may result in a fully automated identification of the disease. Due to the rapid growth of cases every day, a relatively small number of COVID-19 testing kits are readily accessible in health care facilities. Thus it is imperative to define a fully automated detection method as an instant alternate treatment possibility to limit the occurrence of COVID-19 among individuals. In this paper, a two-step Deep learning (DL) architecture has been proposed for COVID-19 diagnosis using CXR. The proposed DL architecture consists of two stages, "feature extraction and classification". The "Multi-Objective Grasshopper Optimization Algorithm (MOGOA)" is presented to optimize the DL network layers; hence, these networks have named as "Multi-COVID-Net". This model classifies the Non-COVID-19, COVID-19, and pneumonia patient images automatically. The Multi-COVID-Net has been tested by utilizing the publicly available datasets, and this model provides the best performance results than other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Unplanned outage due to faults in a high-fidelity turbomachine such as steam turbine and centrifugal compressor often results in the reduced reliability and productivity of a factory while increasing its maintenance costs. Shaft orbit images generated from turbomachine vibration signals have been used to diagnose component faults. However, the existing methods were developed mostly by either using features extracted from orbits or utilizing simulation data which may produce inaccurate results in practical applications due to system complexity and data uncertainties. This paper presents a novel deep learning convolution neural network methodology for accurately automatic diagnostics of multiple faults in general rotating machines by adeptly integrating advanced signal processing with orbit images augmentation, considering the high non-linearity and uncertainty of sensed vibration signals. Environmental noise in vibration signals are filtered through the integration of multiresolution discrete wavelet packet transform and Bayesian hypothesis testing-based automatic thresholding. Shaft orbit images generated from the cleansed vibration data are augmented to increase their representativity and generalization. A novel multi-layer convolutional neural network model, OrbitNet, is specially designed to improve its generality and robustness while avoid possible overfitting in fault identification of various turbomachines. The proposed model retains the pattern information in the axis trajectory to the greatest extent, with the ability of accurately capturing features of various faults in different turbomachines. A generic implementation procedure is proposed for automatic fault diagnosis of rotating machinery based on the presented methodology. A comparison study is conducted to demonstrate the effectiveness and feasibility of the proposed methodology by using the sensed vibration signals collected from three real-world centrifugal compressors, two steam turbines and one generator with four different fault modes including imbalance, friction, misalignment and oil whirl. [Display omitted] • Orbit image-based CNN model for automatic fault diagnostics of turbomachines. • Bayesian wavelets signal processing with sample entropy and image augmentation. • Generalized procedure to implement new methodology for multiple fault diagnostics. • A comparison study on sensor data and different faults from various turbomachines. [ABSTRACT FROM AUTHOR]

Fake news has become a major concern over the Internet. It influences people directly and should be identified. In the recent years, various Machine Learning (ML) and Deep Learning (DL) based data-driven approaches have been suggested for fake news classification. Most of the ML based approaches use hand-crafted features extracted from input textual content. Moreover, in DL based approaches, an efficient word embedding representation of input data is also a major concern. This paper presents a deep learning framework, BerConvoNet , to classify the given news text into fake or real with minimal error. The presented framework has two main building blocks: a news embedding block (NEB) and a multi-scale feature block (MSFB). NEB uses Bidirectional Encoder Representations from Transformers (BERT) for extracting word embeddings from a news article. Next, these embeddings are fed as an input to MSFB. The MSFB consists of multiple kernels (filters) of varying sizes. It extracts various features from news word embedding. The output of MSFB is fed as an input to a fully connected layer for classification. To validate the performance of BerConvoNet , several experiments have been performed on four benchmark datasets and various performance measures are used to evaluate the results. Furthermore, the ablative experiments with respect to news article embedding, kernel size, and batch size have been carried out to ensure the quality of prediction. Comparative analysis of the presented model is done with other state of the art models. It shows that BerConvoNet outplays other models on various performance metrics. • Hybrid deep learning model for textual fake new classification. • Contextual meaning of words in news articles considered. • Multi scale feature extraction using Convolutional Neural Network. • Ablation study of the presented fake news classification model. • Performance Analysis on four benchmark fake news datasets. [ABSTRACT FROM AUTHOR]

In this paper, we address a problem of vision-based 3D robotic navigation using deep reinforcement learning for an Autonomous Underwater Vehicle (AUV). Our research offers conclusions from the experimental study based on one of the RoboSub 2018 competition tasks. However, it can be generalized to any navigation task consisting of movement from a starting point to the front of the next station. The presented reinforcement learning-based model predicts the robot's steering settings using the data acquired from the robot's sensors. Its Vision Module may be based on a built-in convolutional network or a pre-trained TinyYOLO network so that a comparison of various levels of features' complexity is possible. To enable evaluation of the proposed solution, we prepared a test environment imitating the real conditions. It provides the ability to steer the agent simulating the AUV and calculate values of rewards, used for training the model by evaluating its decisions. We study the solution in terms of the reward function form, the model's hyperparameters and the exploited camera images processing method, and provide an analysis of the correctness and speed of the model's functioning. As a result, we obtain a valid model able to steer the robot from the starting point to the destination based on visual cues and inputs from other sensors. [Display omitted] • Examining vision-based, deep reinforcement learning approach for Autonomous Underwater Vehicle (AUV) navigation. • Proposing and analyzing reward function components relevant to a 3D navigation problem. • Researching various image processing methods and visual features' complexity used in the AUV controller. • Evaluating the solution in a realistic simulation environment, regarding its success rate, achieved reward value and prediction speed. [ABSTRACT FROM AUTHOR]

Covid-19 has become a deadly pandemic claiming more than three million lives worldwide. SARS-CoV-2 causes distinct pathomorphological alterations in the respiratory system, thereby acting as a biomarker to aid its diagnosis. A multimodal framework (Ai-CovScan) for Covid-19 detection using breathing sounds, chest X-ray (CXR) images, and rapid antigen test (RAnT) is proposed. Transfer Learning approach using existing deep-learning Convolutional Neural Network (CNN) based on Inception-v3 is combined with Multi-Layered Perceptron (MLP) to develop the CovScanNet model for reducing false-negatives. This model reports a preliminary accuracy of 80% for the breathing sound analysis, and 99.66% Covid-19 detection accuracy for the curated CXR image dataset. Based on Ai-CovScan, a smartphone app is conceptualised as a mass-deployable screening tool, which could alter the course of this pandemic. This app 's deployment could minimise the number of people accessing the limited and expensive confirmatory tests, thereby reducing the burden on the severely stressed healthcare infrastructure. [Display omitted] • This paper proposes Ai-CovScan, a multimodal deep-learning framework for detecting Covid-19. • This framework comprises a transfer-learning-based CovScanNet, where CNN-output is fed into an MLP. • Breathing sound spectrograms are analysed for abnormalities using CovScanNet for Covid-19 prognosis. • A chest X-ray image analysis is also performed for detecting Covid-19 using a curated dataset. • The system is implemented using a smartphone application for rapid detection of Covid-19. [ABSTRACT FROM AUTHOR]

Vehicle detection is an important part of robot environmental perception. In this paper, a 3D vehicle detection method using a single image is proposed to generate the 3D space coordinate information of the object using monocular vision for autonomous driving. The proposed method works under the multi-task framework and integrates 2D object detection, 3D object detection, orientation estimation and key point detection into one unified deep convolution neural network (DCNN) which could be trained by end-to-end learning. Besides, our proposed method is built by modifying Fast R-CNN using multi-task learning, and thus our proposed method is named multi-task Faster R-CNN (MT-Faster R-CNN). The experiments on KITTI dataset are conducted to evaluate our proposed method and the other 3D vehicle detection methods. The experimental results demonstrate that our proposed method is competitive and could significantly assist autonomous driving. • Our proposed MT-Faster R-CNN could be trained by end-to-end learning, which makes it more effective for optimizing. • Our proposed network can simultaneously output 2D and 3D detection results based on single image. • Extensive experiments show that we outperform our baseline with a large margin on KITTI dataset. [ABSTRACT FROM AUTHOR]

Characterization of nucleus from oral tissue histology images is important for oral squamous cell carcinoma (OSCC) diagnosis. Therefore, the accurate delineation of nucleus is the foremost task to initiate the quantification of its various properties such as morphology, texture, intensity, etc. In this paper, we propose a two-stage computer-aided tool for automatic detection (the first stage) and delineation of the detected nucleus (second stage) from oral histological images to aid clinician in OSCC diagnosis. In the first stage, 81 × 81 patches were extracted from our image database followed by the wavelet-based downsampling process to generate 21 × 21 patches, which were used to train the deep convolution neural network (CNN) architecture. This approach was used to automatically detect nucleus from OSCC images and achieved 88.87% recall and 82.03% precision. After detection, 81 × 81 image patches were extracted from those detected nucleus regions and further processed through the second stage of proposed pipeline to segment the exact boundary of nucleus using active contour (Chan-Vese model) after applying edge enhancing nonsubsampled contourlet transform. The proposed segmentation methodology performed well with 94.22% Dice coefficient, 89.38% Jaccard index, 97.56% precision, and 91.58% recall. We view this work as the first attempt on oral tissue histology image computation for joint nucleus detection and segmentation to diagnose OSCC to the best of our knowledge. • Two stage method for nucleus localization and delineation from OSCC histological images. • Computationally efficient deep CNN for nucleus localization. • Nonsubsampled contourlet transform is used to enhance the nucleus border. • Active contour on enhanced images provides improved results of nucleus delineation. [ABSTRACT FROM AUTHOR]