Table 1. Description of some related work indicating pros and cons of existing approaches.
Ref | Data set | Method | pros | cons |
---|---|---|---|---|
Pereira et al. (2020) | RYDLS-20 | Early and late fusion techniques | Multi-class classification+ hierarchical classification | Data sparsity and feature optimization missing |
Apostolopoulos, Aznaouridis & Tzani (2020) | Custom | Mobile Net | Considered feature extraction approaches | 6-class problem |
Oh, Park & Ye (2020) | Custom | Patch-based CNN | Employed Potential imaging biomarkers for validation | 5-class problem |
Apostolopoulos & Mpesiana (2020) | Custom | Transfer learning based CNN | Considered bacterial and viral pneumonia target class | 5-class problem |
Yoo et al. (2020) | Custom | Decision-tree classifier | Merging deep learning in tree classifier | 3-class problem |