Skip to main content
[Preprint]. 2024 Nov 6:2024.11.04.621759. [Version 1] doi: 10.1101/2024.11.04.621759

Figure 1. Zebrafish nnUNET architecture.

Figure 1.

Convolutional blocks within the network contain sequences of a convolution, followed by instance realization and leaky RELU activation function. Block number represents the output size, with downsampling implemented as stridden convolution and upsampling implemented as a convolution transposed.