Table 2.
Summary of papers for pneumonia detection using deep learning
| Reference | Deep Learning Technique | Features | Dataset |
|---|---|---|---|
| [99] | Deep Siamese based neural network | CNN extracted features from the left half and right half of the lungs | Unspecified Kaggle dataset |
| [20] | CNN with transfer learning and data augmentation | Features extracted from CNN | LDOCTCXR |
| [55] | CNN with transfer learning, data augmentation and ensemble by majority voting. | Features extracted from CNN | LDOCTCXR |
| [93] | CNN with transfer learning | Features extracted from CNN | LDOCTCXR |
| [102] | CNN with transfer learning, data augmentation and ensemble by combining confidence scores and bounding boxes. | Features extracted from CNN | Radiological Society of North America (RSNA) pneumonia dataset |
| [96] | CNN with transfer learning and data augmentation | Features extracted from CNN | NIH Chest X-ray Dataset |
| [92] | CNN from scratch and data augmentation | Features extracted from CNN | LDOCTCXR |
| [95] | CNN with transfer learning | Features extracted from CNN | LDOCTCXR |
| [91] | CNN | Features extracted from CNN | Mooney’s Kaggle dataset |
| [100] | CNN and LSTM-CNN, with transfer learning and data augmentation | Features extracted from CNN | Mooney’s Kaggle dataset |
| [103] | CNN with probabilistic map of pneumonia | Features extracted from CNN | 2018 RSNA pneumonia challenge dataset |
| [101] | Decision Tree, Random Forest, K-nearest neighbour, AdaBoost, Gradient Boost, XGBboost, CNN | Multiple features | Mooney’s Kaggle dataset |
| [98] | CNN with transfer learning, data augmentation and ensemble by weighted averaging | Features extracted from CNN | LDOCTCXR |
| [97] | CNN with transfer learning and data augmentation | Features extracted from CNN | Mooney’s Kaggle dataset |
| [94] | CNN with transfer learning | Features extracted from CNN | Private dataset |