Framework for Classification of Chest X-Rays into Normal/COVID-19 Using Brownian-Mayfly-Algorithm Selected Hybrid Features

Table 1

Summary of chosen methodologies employed to detect COVID-19 from X-ray images.

Reference	Methodology employed	Performance metrics (%)
		Accuracy	Sensitivity:	Specificity:
Narin et al. [6]	The performance of pretrained deep-learning scheme supported COVID-19 detection is demonstrated using X-ray images	98.00	—	—
Apostolopoulos and Mpesiana [7]	Convolutional-neural-network (CNN) with transfer learning is employed to examine X-ray to detect COVID-19	93.48	92.85	98.75
Chouhan et al. [8]	Transfer learning based deep-learning scheme is employed to recognize pneumonia in X-rays	96.39	—	—
Stephen et al. [9]	Automatic detection of pneumonia in X-ray is performed using transfer learning	95.00	—	—
Liang and Zheng [10]	Classification of paediatric pneumonia in X-ray is achieved using pretrained CNN	90.00	—	—
Nour et al. [11]	Detection of COVID-19 in X-ray is discussed with deep features and Bayesian optimization	98.97	89.39	99.75
Brunese et al. [12]	Implementation of explainable deep-learning scheme to detect pulmonary abnormality and COVID-19 is presented with X-ray	96.00	96.00	98.00
Ardakani et al. [13]	A detailed analysis of ten widely adopted deep learning methods is discussed and their performance in detection the COVID-19 is demonstrated	99.02	98.04	100
Jaiswal et al. [14]	DenseNet201 supported detection of COVID-19 in X-ray is demonstrated with transfer learning technique	96.25	96.29	96.21
Ucar and Korkmaz [15]	This work proposed deep Bayes-SqueezeNet to detect COVID-19 in X-rays	98.26	99.13	—
Saiz and Barandiaran [16]	This work presented pretrained VGG16 based COVID-19 in X-rays	94.92	94.92	92.00
Panwar et al. [17]	This work demonstrated nCOVnet based identification of COVID-19 in X-rays	88.10	97.62	78.57
Waheed et al. [18]	This work developed a novel deep-learning scheme CovidGAN to detect COVID-19 in X-ray pictures	95.00	90.00	97.00
Kannan et al. [19]	This work demonstrated a study with various pretrained scheme supported COVID-19 classification, and the result of this study confirms that the VGG16 along with K-nearest neighbor (KNN) helped to get better accuracy with Deep + HF	96.48	95.56	95.37