|
Rayan & Alaerjan (2023)
|
X-ray image |
BiLSTM |
161 |
3 |
CV |
T 80%: V 20% |
AC 93.47%, S: 96.15%, SP 90% |
Enhancing Crow Search Optimization through Bi-LSTM Model for SARS-CoV-2 Infection Identification and Classification |
Worked with very small dataset |
|
AlMohimeed et al. (2023)
|
X-ray image |
Stacked RNN |
286, 4,347 |
2 |
CV |
T 80%: V 20% |
AC: 96.83% 98.28% |
Utilized stacked ensemble model using SARS-CoV-2 symptoms and chest X-ray images for the detection of the disease. |
Testing performance is lower comparatively |
|
Aslan et al. (2021)
|
X-ray image |
Stacked RNN |
2,905 |
3 |
CV |
T 80%: V 20% |
AC: 98.14% 98.70% |
BiLSTM used to handle temporal properties of SARS-CoV-2 image and CNN used for the classification. |
Needs to be improve to increase the adaptability to handle new data for automatic detection of SARS-CoV-2. |
|
Muhammad et al. (2022)
|
X-ray image |
CNN BiLSTM |
900, 1,212, 2,020, 2,232 |
2 |
CV |
T 80%: V 20% |
AC: 97%, 84%, 98% |
CNN extracts high-level features from the pooling layer, the augmentation mechanism selects relevant features for low dimensional augmentation, and BiLSTM is employed to classify the processed sequential information. |
Feature extraction and its result visualization needs to be improved. |
|
Demir (2021)
|
X-ray image |
Deep LSTM |
761 |
3 |
CV |
T 80%: V 20% |
AC: 100% |
GRU to extract features from the chest X-ray images, and then uses a CNN layer to classify |
Time and space complexity is higher |
|
Afshar et al. (2020)
|
X ray |
RNN Capsule Network |
13,800 |
3 |
CV |
T 90%: V 10% |
AC: 95.7%% P: 95.8% SN: 90% |
Hybrid mechanism of capsule and RNN gives good outcome |
Less adaptable with big data |
|
Islam et al. (2020)
|
X ray |
CNN- LSTM |
40,000 |
3 |
CV |
T 80%: V 20% |
AC: 99.4% S: 99.2% F1: 98.9% |
Hybrid but nicely handle features and classify it |
Need to be adaptable to analyse multi class data |
|
Ozturk et al. (2020)
|
X-ray |
DarkCovidNet |
1,000 |
2 |
5 Fold |
T 80%: V 20% |
AC: 98.08%, P: 98.03%, R: 95.13%, F1: 96.51% |
Relatively more successful in detection of SARS-CoV-2 |
Less Robust Did not handle real-time data |
|
Shah et al. (2021)
|
X-ray |
CNN-BiGRU |
424 |
3 |
CV |
T 80%: V 20% |
AC: 96.00%, P: 96.00%, R: 96%, F1: 95% |
Relatively successful as a hybrid model in SARS-CoV-2 detection |
Data augmentation problem |
