Table 3.
Summary of feature extraction methods and GI tract disease classification
| Refs | Years | Methods | Datasets | Modality | Results |
|---|---|---|---|---|---|
| [205] | 2023 | Feature Engineering methods are employed for GI tract disease detection and classification. | 8000 images | VE |
99.24% Acc |
| [206] | 2023 | Hybrid approach is used for GI tract disease classification. | 8000 images | VE |
97.25% Acc |
| [207] | 2022 | Deep features are extracted for GI tract disease identification. | 8000 images | VE |
97.00% Acc |
| [29] | 2021 | Transfer learning approaches are used with variants of SVM classifiers GI tract diseases are categorized. | 4000 images | VE |
95.02% Acc |
| [161] | 2020 |
CS-LBP and auto color correlogram are employed for feature extraction and K-mean and SVM classify the frames of endoscopy. |
200 images | WCE |
95.00% Acc |
| [208] | 2020 | GLRLM-based features are used for colorectal polyp findings using SVM. | 86 videos | VE |
98.83% Acc |
| [209] | 2020 | Color and texture features are employed for Polyps identification and classification using an SVM classifier. |
300 Images |
WCE |
86.00% Rec |
| [210] | 2019 | ASWSVD is used for feature extraction and multiple classifiers classify the diseases of GI tract | 5,293 and 8,740 images | WCE | 86.00% Pre |
| [211] | 2019 | An ulcer is classified by an SVM classifier using color and texture features. | 9000 images | WCE |
99.00% Acc |
| [212] | 2018 | Cancer is identified using GLCM and Gabor texture methods and disease classification is performed by multiple classifiers. |
176 Images |
CH |
87.20% Acc |
| [106] | 2018 | Cancer identification and classification are performed by RF and KNN methods using Fourier, HIS, and Statistical techniques for feature extraction. | 280 images | VE | 86.00% Sen |
| [213] | 2017 | Polyps are detected using super pixel-based clustering and SVM classifiers. | 39 images | WCE |
94.00% Acc |