Skip to main content

Table 8 Comparison of the selected deep neural networks in terms of computational complexity, the number of learnable parameters and training/inference speed

From: Semantic segmentation of human oocyte images using deep neural networks

Deep model CCM Total learnables Training-Time Inference-Time
[GFLOPs] [millions] [s/epoch] [ms]
DeepLab...ResNet-18 (15) 3.637 \(\sim \) 17.5 35 36
DeepLab...ResNet-50 (7) 8.2164 \(\sim \) 43 376 105
DeepLab...Incpetion...v2 (10) 22.227 \(\sim \) 71 623 179