Table 4 Comparing the time to identify a single sentiment sample by different network models.

From: Micro-expression recognition model based on TV-L1 optical flow method and improved ShuffleNet

Approaches

Time (ms)

ResNet50

105.9

DeiT39

168

MobileNetV242

24

ShuffleNetV2 + CBAM

28.5

ShuffleNetV2 + ViT

37.9

  1. The test results of our proposed method are shown in bold.
Back to article page