Research Article
CNN Based Multiclass Brain Tumor Detection Using Medical Imaging
Table 1
The summary of model description.
| Layer type | Filter | Kernel size | Output shape | Param# |
| Input layer | — | — | 224 × 224 × 3 | 0- | Convolution | 64 | 3 × 3 | 224 × 224 × 64 | 1792 | Activation | — | — | 224 × 224 × 64 | 0 | BN | — | — | 224 × 224 × 64 | 256 | Convolution | 64 | 3 × 3 | 222 × 222 × 64 | 36928 | Activation | — | — | 222 × 222 × 64 | 0 | Max pooling | 1 | 2 × 2 | 111 × 111 × 64 | 0 | BN | — | — | 111 × 111 × 64 | 256 | Dropout | — | — | 111 × 111 × 64 | 0 | Convolution | 64 | 3 × 3 | 109 × 109 × 64 | 36928 | Activation | — | — | 109 × 109 × 64 | 0 | Max pooling | 1 | 2 × 2 | 54 × 54 × 64 | 0 | BN | — | — | 54 × 54 × 64 | 256 | Dropout | — | — | 54 × 54 × 64 | 0 | Convolution | 64 | 3 × 3 | 54 × 54 × 64 | 36928 | Activation | — | — | 54 × 54 × 64 | 0 | BN | — | — | 54 × 54 × 64 | 256 | Flatten | — | — | 186624 | 0 | Dropout | — | — | 186624 | 0 | FC | — | — | 512 | 95552000 | Activation | — | — | 512 | 0 | BN | — | — | 512 | 2048 | Output layer | — | — | 4 | 2052 | Total params: 95,706,884 | Trainable params: 95,705,220 | Nontrainable params: 1,664 |
|
|