Table 1 3L-CNN configuration for automatic feature extraction
From: Optical electrocardiogram based heart disease prediction using hybrid deep learning
Layers | Convolutional layer | Batch normalization | ReUL layer | Pooling layer |
|---|---|---|---|---|
Layer 1 | Conv1D (1, 64, 40, 4): Input: 1 channels Output: 64 channels kernel size: 40 Stride: 4 | BatchNorma1D (64): Features: 64 | ReUL1D(64): Features: 64 | MaxPool1D(64): kernel size: 4 |
Layer 2 | Conv1D (64, 64, 3, 4): Input: 1 channels Output: 64 channels kernel size: 4 Stride: 3 | BatchNorma1D (64): Features: 64 | ReUL1D(64): Features: 64 | MaxPool1D(64): kernel size: 4 |
Layer 3 | Conv1D (64, 128, 3, 4): Input: 1 channels Output: 128 channels kernel size: 4 Stride: 3 | BatchNorma1D (128): Features: 64 | ReUL1D(128): Features: 128 | MaxPool1D(128): kernel size: 4 |