|
Akl et al. (2023)
|
X-ray image |
Hybrid CNN |
2,482 |
2 |
CV |
T 80%: V 20% |
AC: 99.39% |
Scale-Invariant Feature Transform (SIFT) used for feature extraction and Hybrid deep learning boost detection performance |
Real time detection is not proposed |
|
Kaya & Gürsoy (2023)
|
X-ray image |
MobileNet |
9,457 |
3 |
5k-CV |
T 80%: V 20% |
AC: 95.62%, 96.10%, 97.61% |
Scale-Invariant Feature Transform (SIFT) used for feature extraction and Hybrid deep learning boost detection performance |
Real time detection is not proposed |
|
Kuzinkovas & Clement (2023)
|
X-ray image |
Ensemble CNN |
33,920 |
2 |
CV |
T 80%: V 20% |
AC: 98.34% |
ResNet and VGGG pre-trained models are used for the prediction |
Generalizability is weak |
|
Choudhary et al. (2023)
|
X-ray image |
ResNet |
2,482 |
2 |
CV |
T 80%: V 20% |
AC: 95.47% |
Incorporation of a ResNet34 high resolution network model utilized in SARS-CoV-2 detection |
Did not show real-time image prediction |
|
Khan et al. (2023)
|
X-ray image |
Light CNN |
380 |
2 |
5K-CV |
T 80%: V 20% |
AC: 98.8% |
Novel light CNN model using watershed region-growing segmentation for Chest X-rays image |
Comparative Performance low when working with multiclass data. |
|
Chakraborty, Murali & Mitra (2022)
|
X-ray |
ResNet |
10,040 |
3 |
CV |
T 80%: V 20% |
AC: 96.43% S: 93.68% |
Detect SARS-CoV-2 efficiently |
Lack of model validation Data set is small |
|
Bhattacharyya et al. (2022)
|
X-ray |
VGG-19, CGAN |
3,750 |
3 |
CV |
T 80%: V 20% |
AC: 96.6% |
Automatic SARS-CoV-2 detection with CNN model |
Overfitting and under fitting |
|
Aggarwal et al. (2022)
|
X-ray |
Transfer Learning |
959 |
4 |
CV |
T 80%: V 20% |
AC: 97% |
Efficiently used transfer learning for SARS-CoV-2 analysis |
Learning on small data, Performance low. |
|
Islam et al. (2022)
|
X-ray |
DenseNet |
231 |
3 |
CV |
T 80%: V 20% |
AC: 96.49% |
Properly utilized CNN in COVID recognition. |
Overfitting in performance and used very Small data |
|
Jain et al. (2021)
|
X-ray |
Inception V3/Xception Net/ResNet |
6,432 |
3 |
CV |
T 80%: V 20% |
P: 99% R: 92% F1: 95% |
Identify SARS-CoV-2 by CNN properly |
Under fitting problem Small data |
|
Zebin & Rezvy (2021)
|
X-ray |
VGG16, ResNet50, Efficient Net |
822 |
3 |
5 Fold |
T 80%: V 20% |
AC: 90.0% AC: 94.3% AC: 96.8% |
Nicely Visualize infected lung area to detect SARS-CoV-2 |
Operational complexity is slightly higher |
|
Sait et al. (2021)
|
Sound, Image |
Inception |
4,558 |
2 |
CV |
T 80%: V 20% |
AC: 80% AC: 99.66% |
High performed Multi-modal for SARS-CoV-2 analysis |
Time complexity higher and Limited data |
