Research Article
Security and Privacy of Cloud- and IoT-Based Medical Image Diagnosis Using Fuzzy Convolutional Neural Network
| Algorithm | Parameters | Values |
| KNN | Number of neighbours | 5 | Distance function | Euclidean distance | (N × D) training data | N, no. of samples; D, dimensionality of each data point | (M × D) testing data | M, no. of data points | NB | Model | Gaussian base distribution | N | Size of data | DT | Splitting criterion | Gini | Minimum instances per leaf | 2 | ANN | Size of input layer | Size of data | Type of ANN | Feed-forward | Number of neurons | 20 | Training and testing set | 75% of training and 25% of testing set | FCNN | Input | 56 × 28 | Fuzzification | 2 × (input)-Gaussian MF | In and out channel range | 1 to 100 | Stride and padding | 1 & 0 | Conv3x d | 2 × (in & out channels, kernel size = (3 × 128), stride & padding), ReLU, Max_Pooling (55 × 1) | Conv4x d | 2 × (in & out channels, kernel size = (4 × 128), stride & padding), ReLU, Max_Pooling (54 × 1) | Conv5x d | 2 × (in & out channels, kernel size = (5 × 128), stride & padding), ReLU, Max_Pooling (53 × 1) | Defuzzification | 2 × 128 |
|
|