Table 2 Total number parameters for the proposed CNN with a channel attention-based model.
From: Stacked CNN-based multichannel attention networks for Alzheimer disease detection
Type of Layer | Output | Number of Parameter |
|---|---|---|
Input Layer | (None, M, N, 3) | 0 |
(CNNB-1) | (None, M, N, 16) | 1216 |
(CNNB-2) | (None, 82, 98, 32) | 12832 |
(CNNB-3) | (None, 37, 45, 64) | 37264 |
(CNNB-4) | (None, 14, 18, 128) | 204928 |
(CNNB-5) | (None, 3, 5, 256) | 819456 |
Channel-attention | (None, 1, 2, 256) | 33088 |
(Dropout-1) | (None, 1, 2, 256) | 0 |
(Flatten) | (None, 512) | 0 |
(Dense-1) | (None, 256) | 131328 |
(Dropout-2) | (None, 256) | 0 |
(Dense-2) | (None, 4) | 1028 |
(Soft-max) | (None, 4) | 0 |
Total params: | 1,241,140 | |
Train params: | 1,241,140 | |
Non-train params: | 0 |
