SharifRazavian, A., Azizpour, H., Sullivan, J. According to the promising results of the proposed model, that combines CNN as a feature extractor and FO-MPA as a feature selector could be useful and might be successful in being applied in other image classification tasks. Faramarzi et al.37 divided the agents for two halves and formulated Eqs. Pool layers are used mainly to reduce the inputs size, which accelerates the computation as well. Both the model uses Lungs CT Scan images to classify the covid-19. Adv. The survey asked participants to broadly classify the findings of each chest CT into one of the four RSNA COVID-19 imaging categories, then select which imaging features led to their categorization. Chollet, F. Keras, a python deep learning library. Based on54, the later step reduces the memory requirements, and improve the efficiency of the framework. Can ai help in screening viral and covid-19 pneumonia? Detection of lung cancer on chest ct images using minimum redundancy maximum relevance feature selection method with convolutional neural networks. The results of max measure (as in Eq. Cancer 48, 441446 (2012). org (2015). For the image features that led to categorization, there were varied levels of agreement in the interobserver and intraobserver components that . For example, as our input image has the shape \(224 \times 224 \times 3\), Nasnet26 produces 487 K features, Resnet25 and Xception29 produce about 100 K features and Mobilenet27 produces 50 K features, while FO-MPA produces 130 and 86 features for both dataset1 and dataset 2, respectively. Number of extracted feature and classification accuracy by FO-MPA compared to other CNNs on dataset 1 (left) and on dataset 2 (right). Future Gener. Coronavirus disease (Covid-19) is an infectious disease that attacks the respiratory area caused by the severe acute . 40, 2339 (2020). This dataset consists of 219 COVID-19 positive images and 1341 negative COVID-19 images. Diagnosis of parkinsons disease with a hybrid feature selection algorithm based on a discrete artificial bee colony. In the last two decades, two famous types of coronaviruses SARS-CoV and MERS-CoV had been reported in 2003 and 2012, in China, and Saudi Arabia, respectively3. In Eq. Kharrat, A. Use the Previous and Next buttons to navigate the slides or the slide controller buttons at the end to navigate through each slide. The variants of concern are Alpha, Beta, Gamma, and than the COVID-19 images. To further analyze the proposed algorithm, we evaluate the selected features by FO-MPA by performing classification. Inspired by this concept, Faramarzi et al.37 developed the MPA algorithm by considering both of a predator a prey as solutions. 101, 646667 (2019). In this paper, each feature selection algorithm were exposed to select the produced feature vector from Inception aiming at selecting only the most relevant features. Efficient classification of white blood cell leukemia with improved swarm optimization of deep features. Furthermore, deep learning using CNN is considered one of the best choices in medical imaging applications20, especially classification. COVID-19 image classification using deep features and fractional-order marine predators algorithm Authors. arXiv preprint arXiv:2003.13145 (2020). On January 20, 2023, Japanese Prime Minister Fumio Kishida announced that the country would be downgrading the COVID-19 classification. Decaf: A deep convolutional activation feature for generic visual recognition. This stage can be mathematically implemented as below: In Eq. J. Med. CAS Some people say that the virus of COVID-19 is. It based on using a deep convolutional neural network (Inception) for extracting features from COVID-19 images, then filtering the resulting features using Marine Predators Algorithm (MPA), enhanced by fractional-order calculus(FO). Comput. In some cases (as exists in this work), the dataset is limited, so it is not sufficient for building & training a CNN. & Pouladian, M. Feature selection for contour-based tuberculosis detection from chest x-ray images. It is obvious that such a combination between deep features and a feature selection algorithm can be efficient in several image classification tasks. Sci. Table4 show classification accuracy of FO-MPA compared to other feature selection algorithms, where the best, mean, and STD for classification accuracy were calculated for each one, besides time consumption and the number of selected features (SF). In this paper, we propose an improved hybrid classification approach for COVID-19 images by combining the strengths of CNNs (using a powerful architecture called Inception) to extract features and a swarm-based feature selection algorithm (Marine Predators Algorithm) to select the most relevant features. The parameters of each algorithm are set according to the default values. Therefore, a feature selection technique can be applied to perform this task by removing those irrelevant features. Furthermore, using few hundreds of images to build then train Inception is considered challenging because deep neural networks need large images numbers to work efficiently and produce efficient features. COVID-19-X-Ray-Classification Utilizing Deep Learning to detect COVID-19 and Viral Pneumonia from x-ray images Research Publication: https://dl.acm.org/doi/10.1145/3431804 Datasets used: COVID-19 Radiography Database COVID-19 10000 Images Related Research Papers: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7187882/ They employed partial differential equations for extracting texture features of medical images. Figure7 shows the most recent published works as in54,55,56,57 and44 on both dataset 1 and dataset 2. Zhang, N., Ruan, S., Lebonvallet, S., Liao, Q. A deep feature learning model for pneumonia detection applying a combination of mRMR feature selection and machine learning models. Havaei, M. et al. COVID-19 (coronavirus disease 2019) is a new viral infection disease that is widely spread worldwide. Li et al.36 proposed an FS method using a discrete artificial bee colony (ABC) to improve the classification of Parkinsons disease. Sci Rep 10, 15364 (2020). Book Google Scholar. Based on Standard Deviation measure (STD), the most stable algorithms were SCA, SGA, BPSO, and bGWO, respectively. 1. Hence, it was discovered that the VGG-16 based DTL model classified COVID-19 better than the VGG-19 based DTL model. 69, 4661 (2014). Lilang Zheng, Jiaxuan Fang, Xiaorun Tang, Hanzhang Li, Jiaxin Fan, Tianyi Wang, Rui Zhou, Zhaoyan Yan: PVT-COV19D: COVID-19 Detection Through Medical Image Classification Based on Pyramid Vision Transformer. In Smart Intelligent Computing and Applications, 305313 (Springer, 2019). In Future of Information and Communication Conference, 604620 (Springer, 2020). In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 770778 (2016). This dataset currently contains hundreds of frontal view X-rays and is the largest public resource for COVID-19 image and prognostic data, making it a necessary resource to develop and evaluate tools to aid in the treatment of CO VID-19. In this paper, we propose an improved hybrid classification approach for COVID-19 images by combining the strengths of CNNs (using a powerful architecture called Inception) to extract features and . J. Med. Moreover, we design a weighted supervised loss that assigns higher weight for . The main purpose of Conv. For fair comparison, each algorithms was performed (run) 25 times to produce statistically stable results.The results are listed in Tables3 and4. Transmission scenarios for middle east respiratory syndrome coronavirus (mers-cov) and how to tell them apart. Imag. A properly trained CNN requires a lot of data and CPU/GPU time. Med. In our experiment, we randomly split the data into 70%, 10%, and 20% for training, validation, and testing sets, respectively. Recombinant: A process in which the genomes of two SARS-CoV-2 variants (that have infected a person at the same time) combine during the viral replication process to form a new variant that is different . Biol. The \(\delta\) symbol refers to the derivative order coefficient. Image Underst. According to the formula10, the initial locations of the prey and predator can be defined as below: where the Elite matrix refers to the fittest predators. However, it has some limitations that affect its quality. However, the proposed FO-MPA approach has an advantage in performance compared to other works. Comput. \(Fit_i\) denotes a fitness function value. Propose a novel robust optimizer called Fractional-order Marine Predators Algorithm (FO-MPA) to select efficiently the huge feature vector produced from the CNN. Stage 3: This stage executed on the last third of the iteration numbers (\(t>\frac{2}{3}t_{max}\)) where based on the following formula: Eddy formation and Fish Aggregating Devices effect: Faramarzi et al.37 considered the external impacts from the environment, such as the eddy formation or Fish Aggregating Devices (FADs) effects to avoid the local optimum solutions. where \(fi_{i}\) represents the importance of feature I, while \(ni_{j}\) refers to the importance of node j. Apostolopoulos, I. D. & Mpesiana, T. A. Covid-19: automatic detection from x-ray images utilizing transfer learning with convolutional neural networks. As seen in Table1, we keep the last concatenation layer which contains the extracted features, so we removed the top layers such as the Flatten, Drop out and the Dense layers which the later performs classification (named as FC layer). Then, applying the FO-MPA to select the relevant features from the images. Jcs: An explainable covid-19 diagnosis system by joint classification and segmentation. 4a, the SMA was considered as the fastest algorithm among all algorithms followed by BPSO, FO-MPA, and HHO, respectively, while MPA was the slowest algorithm. Extensive evaluation experiments had been carried out with a collection of two public X-ray images datasets. Future Gener. AMERICAN JOURNAL OF EMERGENCY MEDICINE COVID-19: Facemask use prevalence in international airports in Asia, Europe and the Americas, March 2020 One of the best methods of detecting. Recently, a combination between the fractional calculus tool and the meta-heuristics opens new doors in providing robust and reliable variants41. Phys. Layers are applied to extract different types of features such as edges, texture, colors, and high-lighted patterns from the images. The results showed that the proposed approach showed better performances in both classification accuracy and the number of extracted features that positively affect resource consumption and storage efficiency. Multimedia Tools Appl. 97, 849872 (2019). The test accuracy obtained for the model was 98%. Etymology. Article The largest features were selected by SMA and SGA, respectively. Health Inf. We adopt a special type of CNN called a pre-trained model where the network is previously trained on the ImageNet dataset, which contains millions of variety of images (animal, plants, transports, objects,..) on 1000 classe categories. An image segmentation approach based on fuzzy c-means and dynamic particle swarm optimization algorithm. Article (22) can be written as follows: By taking into account the early mentioned relation in Eq. We have used RMSprop optimizer for weight updates, cross entropy loss function and selected learning rate as 0.0001. kharrat and Mahmoud32proposed an FS method based on a hybrid of Simulated Annealing (SA) and GA to classify brain tumors using MRI. Initialize solutions for the prey and predator. 51, 810820 (2011). Moreover, from Table4, it can be seen that the proposed FO-MPA provides better results in terms of F-Score, as it has the highest value in datatset1 and datatset2 which are 0.9821 and 0.99079, respectively. To survey the hypothesis accuracy of the models. Sahlol, A. T., Kollmannsberger, P. & Ewees, A. Kong, Y., Deng, Y. 152, 113377 (2020). With accounting the first four previous events (\(m=4\)) from the memory data with derivative order \(\delta\), the position of prey can be modified as follow; Second: Adjusting \(R_B\) random parameter based on weibull distribution. 115, 256269 (2011). Comput. Podlubny, I. The first one is based on Python, where the deep neural network architecture (Inception) was built and the feature extraction part was performed. The experimental results and comparisons with other works are presented inResults and discussion section, while they are discussed in Discussion section Finally, the conclusion is described in Conclusion section. In Dataset 2, FO-MPA also is reported as the highest classification accuracy with the best and mean measures followed by the BPSO. Detecting COVID-19 at an early stage is essential to reduce the mortality risk of the patients. The convergence behaviour of FO-MPA was evaluated over 25 independent runs and compared to other algorithms, where the x-axis and the y-axis represent the iterations and the fitness value, respectively. In transfer learning, a CNN which was previously trained on a large & diverse image dataset can be applied to perform a specific classification task by23. After applying this technique, the feature vector is minimized from 2000 to 459 and from 2000 to 462 for Dataset1 and Dataset 2, respectively. used VGG16 to classify Covid-19 and achieved good results with an accuracy of 86% [ 22 ]. Also, it has killed more than 376,000 (up to 2 June 2020) [Coronavirus disease (COVID-2019) situation reports: (https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports/)]. 43, 635 (2020). Although outbreaks of SARS and MERS had confirmed human to human transmission3, they had not the same spread speed and infection power of the new coronavirus (COVID-19). The conference was held virtually due to the COVID-19 pandemic. Howard, A.G. etal. & Dai, Q. Discriminative clustering and feature selection for brain mri segmentation. Also, some image transformations were applied, such as rotation, horizontal flip, and scaling. 121, 103792 (2020). PubMedGoogle Scholar. Med. Med. arXiv preprint arXiv:1711.05225 (2017). Epub 2022 Mar 3. In this paper, we try to integrate deep transfer-learning-based methods, along with a convolutional block attention module (CBAM), to focus on the relevant portion of the feature maps to conduct an image-based classification of human monkeypox disease. PubMed Classification of COVID19 using Chest X-ray Images in Keras 4.6 33 ratings Share Offered By In this Guided Project, you will: Learn to Build and Train the Convolutional Neural Network using Keras with Tensorflow as Backend Learn to Visualize Data in Matplotlib Learn to make use of the Trained Model to Predict on a New Set of Data 2 hours This means we can use pre-trained model weights, leveraging all of the time and data it took for training the convolutional part, and just remove the FCN layer. & Cao, J. In the meantime, to ensure continued support, we are displaying the site without styles The proposed CNN architecture for Task 2 consists of 14 weighted layers, in which there are three convolutional layers and one fully connected layer, as shown in Fig. Syst. It is calculated between each feature for all classes, as in Eq. Methods: We employed a public dataset acquired from 20 COVID-19 patients, which . IEEE Trans. Eng. Eng. A. In this subsection, a comparison with relevant works is discussed. }, \end{aligned}$$, $$\begin{aligned} D^{\delta }[U(t)]=\frac{1}{T^\delta }\sum _{k=0}^{m} \frac{(-1)^k\Gamma (\delta +1)U(t-kT)}{\Gamma (k+1)\Gamma (\delta -k+1)} \end{aligned}$$, $$\begin{aligned} D^1[U(t)]=U(t+1)-U(t) \end{aligned}$$, $$\begin{aligned} U=Lower+rand_1\times (Upper - Lower ) \end{aligned}$$, $$\begin{aligned} Elite=\left[ \begin{array}{cccc} U_{11}^1&{}U_{12}^1&{}\ldots &{}U_{1d}^1\\ U_{21}^1&{}U_{22}^1&{}\ldots &{}U_{2d}^1\\ \ldots &{}\ldots &{}\ldots &{}\ldots \\ U_{n1}^1&{}U_{n2}^1&{}\ldots &{}U_{nd}^1\\ \end{array}\right] , \, U=\left[ \begin{array}{cccc} U_{11}&{}U_{12}&{}\ldots &{}U_{1d}\\ U_{21}&{}U_{22}&{}\ldots &{}U_{2d}\\ \ldots &{}\ldots &{}\ldots &{}\ldots \\ U_{n1}&{}U_{n2}&{}\ldots &{}U_{nd}\\ \end{array}\right] , \, \end{aligned}$$, $$\begin{aligned} S_i&= {} R_B \bigotimes (Elite_i-R_B\bigotimes U_i), i=1,2,\ldots ,n \end{aligned}$$, $$\begin{aligned} U_i&= {} U_i+P.R\bigotimes S_i \end{aligned}$$, \(\frac{1}{3}t_{max}< t< \frac{2}{3}t_{max}\), $$\begin{aligned} S_i&= {} R_L \bigotimes (Elite_i-R_L\bigotimes U_i), i=1,2,\ldots ,n/2 \end{aligned}$$, $$\begin{aligned} S_i&= {} R_B \bigotimes (R_B \bigotimes Elite_i- U_i), i=1,2,\ldots ,n/2 \end{aligned}$$, $$\begin{aligned} U_i&= {} Elite_i+P.CF\bigotimes S_i,\, CF= \left( 1-\frac{t}{t_{max}} \right) ^{\left(2\frac{t}{t_{max}}\right) } \end{aligned}$$, $$\begin{aligned} S_i&= {} R_L \bigotimes (R_L \bigotimes Elite_i- U_i), i=1,2,\ldots ,n \end{aligned}$$, $$\begin{aligned} U_i&= {} Elite_i+P.CF\bigotimes S_i,\, CF= \left( 1-\frac{t}{t_{max}}\right) ^{\left(2\frac{t}{t_{max}} \right) } \end{aligned}$$, $$\begin{aligned} U_i=\left\{ \begin{array}{ll} U_i+CF [U_{min}+R \bigotimes (U_{max}-U_{min})]\bigotimes W &{} r_5 < FAD \\ U_i+[FAD(1-r)+r](U_{r1}-U_{r2}) &{} r_5 > FAD\\ \end{array}\right. Duan et al.13 applied the Gaussian mixture model (GMM) to extract features from pulmonary nodules from CT images. Stage 2: The prey/predator in this stage begin exploiting the best location that detects for their foods. Toaar, M., Ergen, B. In my thesis project, I developed an image classification model to detect COVID-19 on chest X-ray medical data using deep learning models such . JMIR Formative Research - Classifying COVID-19 Patients From Chest X-ray Images Using Hybrid Machine Learning Techniques: Development and Evaluation Published on 28.2.2023 in Vol 7 (2023) Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/42324, first published August 31, 2022 . Zhang et al.16 proposed a kernel feature selection method to segment brain tumors from MRI images. These images have been further used for the classification of COVID-19 and non-COVID-19 images using ResNet50 and AlexNet convolutional neural network (CNN) models. In this paper, Inception is applied as a feature extractor, where the input image shape is (229, 229, 3). & Cmert, Z. In14, the authors proposed an FS method based on a convolutional neural network (CNN) to detect pneumonia from lung X-ray images. Both datasets shared some characteristics regarding the collecting sources. The COVID-19 pandemic has been having a severe and catastrophic effect on humankind and is being considered the most crucial health calamity of the century. Computer Department, Damietta University, Damietta, Egypt, Electrical Engineering Department, Faculty of Engineering, Fayoum University, Fayoum, Egypt, State Key Laboratory for Information Engineering in Surveying, Mapping, and Remote Sensing, Wuhan University, Wuhan, China, Department of Applied Informatics, Vytautas Magnus University, Kaunas, Lithuania, Department of Mathematics, Faculty of Science, Zagazig University, Zagazig, Egypt, School of Computer Science and Robotics, Tomsk Polytechnic University, Tomsk, Russia, You can also search for this author in They were manually aggregated from various web based repositories into a machine learning (ML) friendly format with accompanying data loader code. Cohen, J.P., Morrison, P. & Dao, L. Covid-19 image data collection. Thereafter, the FO-MPA parameters are applied to update the solutions of the current population. Toaar, M., Ergen, B. A features extraction method using the Histogram of Oriented Gradients (HOG) algorithm and the Linear Support Vector Machine (SVM), K-Nearest Neighbor (KNN) Medium and Decision Tree (DT) Coarse Tree classification methods can be used in the diagnosis of Covid-19 disease. Eur. Therefore, reducing the size of the feature from about 51 K as extracted by deep neural networks (Inception) to be 128.5 and 86 in dataset 1 and dataset 2, respectively, after applying FO-MPA algorithm while increasing the general performance can be considered as a good achievement as a machine learning goal. Negative COVID-19 images were collected from another Chest X-ray Kaggle published dataset43. However, WOA showed the worst performances in these measures; which means that if it is run in the same conditions several times, the same results will be obtained. Memory FC prospective concept (left) and weibull distribution (right). Eng. 92, 103662. https://doi.org/10.1016/j.engappai.2020.103662 (2020). It classifies the chest X-ray images into three categories that includes Covid-19, Pneumonia and normal. Background The large volume and suboptimal image quality of portable chest X-rays (CXRs) as a result of the COVID-19 pandemic could post significant challenges for radiologists and frontline physicians. Syst. Computed tomography (CT) and magnetic resonance imaging (MRI) represent valuable input to AI algorithms, scanning human body sections for the sake of diagnosis. There are three main parameters for pooling, Filter size, Stride, and Max pool. Future Gener. implemented the FO-MPA swarm optimization and prepared the related figures and manuscript text. Fung, G. & Stoeckel, J. Svm feature selection for classification of spect images of alzheimers disease using spatial information. Dual feature selection and rebalancing strategy using metaheuristic optimization algorithms in x-ray image datasets. For the exploration stage, the weibull distribution has been applied rather than Brownian to bost the performance of the predator in stage 2 and the prey velocity in stage 1 based on the following formula: Where k, and \(\zeta\) are the scale and shape parameters. How- individual class performance. Int. Math. Artif. Shi, H., Li, H., Zhang, D., Cheng, C. & Cao, X. The evaluation outcomes demonstrate that ABC enhanced precision, and also it reduced the size of the features. 6, right), our approach still provides an overall accuracy of 99.68%, putting it first with a slight advantage over MobileNet (99.67 %). We do not present a usable clinical tool for COVID-19 diagnosis, but offer a new, efficient approach to optimize deep learning-based architectures for medical image classification purposes. So, transfer learning is applied by transferring weights that were already learned and reserved into the structure of the pre-trained model, such as Inception, in this paper. Acharya et al.11 applied different FS methods to classify Alzheimers disease using MRI images. They showed that analyzing image features resulted in more information that improved medical imaging.