Table 1 2D-CNN model layers and parameters.
From: ECG heartbeat classification using progressive moving average transform
No. | Layer name | Kernal size | Filter | Padding | Stride | Output shape |
|---|---|---|---|---|---|---|
1 | \(\hbox {Input}^a\) | – | – | – | – | \(120\times 120\times 1\) |
2 | Conv2D | \(5\times 5\) | 32 | 0 | 1 | \(116\times 116\times 32\) |
3 | BatchNorm2d | – | – | – | – | \(116\times 116\times 32\) |
4 | ReLU | – | – | – | – | \(116\times 116\times 32\) |
5 | MaxPool2D | \(5\times 5\) | – | – | 5 | \(23\times 23\times 32\) |
6 | Conv2D | \(5\times 5\) | 64 | 0 | 1 | \(19\times 19\times 64\) |
7 | BatchNorm2d | – | – | – | – | \(19\times 19\times 64\) |
8 | ReLU | – | – | – | – | \(19\times 19\times 64\) |
9 | MaxPool2D | \(3\times 3\) | – | – | 3 | \(6\times 6\times 64\) |
10 | Conv2D | \(3\times 3\) | 128 | 0 | 1 | \(4\times 4\times 128\) |
11 | BatchNorm2d | – | – | – | – | \(4\times 4\times 128\) |
12 | ReLU | – | – | – | – | \(4\times 4\times 128\) |
13 | AMaxPool2\(\hbox {d}^b\) | – | – | – | – | \(1\times 1\times 128\) |
14 | Flatten | – | – | – | – | \(1\times 128\) |
15 | Flatten | – | – | – | – | \(1\times 128\) |
16 | Flatten | – | – | – | – | \(1\times 64\) |
17 | Flatten | – | – | – | – | \(1\times 3\) |
