TABLE 4.
Comparative analysis of proposed approach in terms of computational complexity with the base framework.
Method | Total trainable framework parameters (Million) | Processing time (ms) |
Inception V4 | 41.2 | 4042 |
VGG-16 | 119.6 | 1051 |
ResNet-50 | 23.6 | 1583 |
ResNet-101 | 42.5 | 2766 |
ResNet-152 | 58.5 | 4366 |
DenseNet-201 | 20 | 2573 |
EfficientNet | 19.4 | 1548 |
EfficientNetV2 | 15.2 | 1125 |
Improved EfficientNetV2 | 14.4 | 1053 |