Table 1 Merits and demerits of traditional BC detection methods.
From: Intelligent breast cancer diagnosis with two-stage using mammogram images
Author [citation] | Methodology | Features | Challenges |
|---|---|---|---|
Das et al.20 | CNN and EWT | It improves the detection rate regarding the accuracy, recall, precision, etc | It causes hardware complexity as well as time complexity |
Saber et al.21 | VGG | It achieves more accuracy, AUC, and sensitivity, enhancing the system's robustness | It is further developing for prognosis objectives |
Jiang et al.22 | PAA | It extracts the peripheral regions to get the features for identifying the diseases | It causes a computational burden that degrades the robustness of the system |
Kavitha et al.23 | BPNN | It offers discriminative features for reaching a higher detection rate | It does not tune the parameters used in the model for further enhancement |
Kumari and Jagadesh24 | XGBoost | It chooses the noteworthy features for increasing the detection accuracy | It does not support large-scale dimensional datasets |
Patil and Biradar25 | CNN | It significantly extracts the boundary-level regions for estimating the appropriate results | The blur or noise present in images degrades the system's robustness and misdiagnoses the disease |
Pramanik et al.29 | KNN | It acquires deep and optimal features for detecting the cancer regions in images | Time complexity and premature convergence rate occur |
Zheng et al.30 | CNN-LSTM | It obtains the desired value to detect the disease at its early stages | It causes the overfitting problem |
