Skip to main content
. 2022 Jan 24;16(6):1455–1462. doi: 10.1007/s11760-021-02098-8

Table 3.

Comparison between the proposed 3-stream fusion-based CNN model and the corresponding 1-stream versions

3-Stream network 1-Stream networks
grayscale LBP HOG
Training time (min) 3.55 1.56 1.52 1.80
Prediction time (sec/image) 0.0029 0.0011 0.0011 0.0011
Loss rate 0.07 0.11 0.11 0.12
Accuracy rate 97.76 95.80 95.67 95.31