Skip to main content
. 2023 Feb 1;14:1096802. doi: 10.3389/fpls.2023.1096802

Table 5.

The inference time of the convolutional neural networks evaluated in the study.

Deep learning architecture Image type Resolution Batch size FPS
DenseNet Sub-image 240 × 216 1 103.75
40 61.79
EfficientNet-v2 Sub-image 240 × 216 1 77.94
40 38.77
ResNet Sub-image 240 × 216 1 276.08
40 93.38

FPS, frames per second.