Table 1: Architecture of DenseU-Net, where Dense-1/32 represents a Dense block structure, and its growth rate is 32.

EncoderOutput sizeSkip connectionDecoderOutput size

Input2562 × 1Conv62562 × 1
Dense-1/322562 × 64Res1Conv52562 × 2
Pooling1282 × 64Up42562 × 64
Dense-2/321282 × 128Res2Conv41282 × 64
Pooling642 × 128Up31282 × 128
Dense-2/64642 × 256Res3Conv3642 × 128
Pooling322 × 256Up2642 × 256
Dense-4/64322 × 512Res4Conv2322 × 256
Pooling162 × 512Up1322 × 512
Dense-4/128162 × 1024Conv1162 × 512

Pooling indicates that the maximum pooling layer has a pooling window of 22 and a step size of 2. Conv represents the convolutional layer. Up indicates the upsampling layer. Res stands for Residual block.