Skip to main content
. 2023 May 28;11(6):1566. doi: 10.3390/biomedicines11061566

Figure 1.

Figure 1

Workflow diagram; Stage-1 (Dataset Preparation): Combination of five different datasets; Stage-2: Image pre-processing (A). Artifacts removal, (B). Noise Removal, and (C). Image Enhancement; Stage-3 (Data Balancing): Down sampling (Applying 12 types of augmentation techniques); Stage-4 (Base Model Creation): Apply seven types of CNN models (Compact Convolutional Transformer (CCT), Swin Transformer, Vision Transformer, VGG16, VGG19, MobileNetV2, and InceptionV3); Stage-5 (Ablation Study): Apply twelve types of alteration to the base model for creating a proposed model; Stage-6 (Results and Discussion): Results Analysis, Comparison with Existing Literature, model performance evaluation, and Robustness Analysis.