ChestCovidNet: An Effective DL-based Approach for COVID-19, Lung Opacity, and Pneumonia Detection Using Chest Radiographs Images DOI Open Access
Naeem Ullah, Javed Ali Khan, Sultan Almakdi

et al.

Biochemistry and Cell Biology, Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 2, 2024

Currently used lung disease screening tools are expensive in terms of money and time. Therefore, chest radiograph images (CRIs) employed for prompt accurate COVID-19 identification. Recently, many researchers have applied Deep learning (DL) based models to detect automatically. However, their model could been more computationally less robust, i.e., its performance degrades when evaluated on other datasets. This study proposes a trustworthy, lightweight network (ChestCovidNet) that can by examining various CRIs The ChestCovidNet has only 11 learned layers, eight convolutional (Conv) three fully connected (FC) layers. framework employs both the Conv group Leaky Relu activation function, shufflenet unit, kernels 3×3 1×1 extract features at different scales, two normalization procedures cross-channel batch normalization. We 9013 training whereas 3863 testing proposed approach. Furthermore, we compared classification results with hybrid methods which DL frameworks feature extraction support vector machines (SVM) classification. study's findings demonstrated embedded low-power worked well achieved accuracy 98.12% recall, F1-score, precision 95.75%.

Language: Английский

ChestCovidNet: An Effective DL-based Approach for COVID-19, Lung Opacity, and Pneumonia Detection Using Chest Radiographs Images DOI Open Access
Naeem Ullah, Javed Ali Khan, Sultan Almakdi

et al.

Biochemistry and Cell Biology, Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 2, 2024

Currently used lung disease screening tools are expensive in terms of money and time. Therefore, chest radiograph images (CRIs) employed for prompt accurate COVID-19 identification. Recently, many researchers have applied Deep learning (DL) based models to detect automatically. However, their model could been more computationally less robust, i.e., its performance degrades when evaluated on other datasets. This study proposes a trustworthy, lightweight network (ChestCovidNet) that can by examining various CRIs The ChestCovidNet has only 11 learned layers, eight convolutional (Conv) three fully connected (FC) layers. framework employs both the Conv group Leaky Relu activation function, shufflenet unit, kernels 3×3 1×1 extract features at different scales, two normalization procedures cross-channel batch normalization. We 9013 training whereas 3863 testing proposed approach. Furthermore, we compared classification results with hybrid methods which DL frameworks feature extraction support vector machines (SVM) classification. study's findings demonstrated embedded low-power worked well achieved accuracy 98.12% recall, F1-score, precision 95.75%.

Language: Английский

Citations

8