Table 2 Related methods occupied with the pathological assessment of colorectal polyps
From: Automated classification of polyps using deep learning architectures and few-shot learning
Author | Year | Method | Data | Classification | Accuracy |
|---|---|---|---|---|---|
Ribeiro et al. [24] | 2016 | custom CNN | Private | Healthy abnormal | 90.96 % |
Zhang et al. [26] | 2016 | CaffeNet | Private and [31] | hyperplastic adenoma | 85.9 % |
Bryne et al. [27] | 2017 | InceptionNet | Private | Hyperplastic adenoma | 94 % |
Komeda et al. [28] | 2017 | custom CNN | Private | Adenoma non-adenoma | 75.1 % |
Lui et al. [6] | 2019 | custom CNN | Private | Curable non-curable | 85.5 % |
Bour et al. [4] | 2019 | ResNet-50 | Private | Not dangerous dangerous cancer | 87.1 % |
Tanwar et al. [25] | 2020 | VGG-16 | Private | Benign Malignant Nonmalignant | 84 % |
Ozawa et al. [5] | 2020 | SSD | private | Hyperplastic adenoma | 83 % |
Hsu et al. [29] | 2021 | custom CNN | Private | Hyperplastic neoplastic | 72.2 % (Weight light) 82.8 % (NBI light) |
Chung-Ming et al. [30] | 2022 | AlexNet | Private | Hyperplastic adenoma | 96.4 % |