Research Article
Image Classification Based on Light Convolutional Neural Network Using Pulse Couple Neural Network
Table 8
ImageNet performance metrics comparison.
| Model | Accuracy (top-1) | Accuracy (top-2) | Parameters | Time (ms) per inference step (CPU) | Time (ms) per inference step (GPU) |
| Xception [12] | 0.7900 | 0.9450 | 22.9 M | 109.4 | 8.1 | VGG16 [12] | 0.7130 | 0.9010 | 138.4 M | 69.5 | 4.2 | VGG19 [19] | 0.7130 | 0.9000 | 143.7 M | 84.8 | 4.4 | ResNet152V2 [20] | 0.7800 | 0.9420 | 60.4 M | 107.5 | 6.6 | MobileNetV2 [20] | 0.7130 | 0.9010 | 3.5 M | 25.9 | 3.8 | NASNetLarge [21] | 0.8250 | 0.9600 | 88.9 M | 344.5 | 20.0 | EfficientNetB7 [21] | 0.8430 | 0.9700 | 66.7 M | 1578.9 | 61.6 | Proposed method | 0.9199 | 0.9573 | 201019 | 25.1 | 3.0 |
|
|