MPU-Net

The code is for the article 3D Medical Image Segmentation based on multi-scale MP-UNet

Structure

• Contraction Phase (Encoder):

  • The input image is passed through a Convolutional Neural Network (CNN) layer to extract high-level feature representations.
  • The CNN's output is then serialized, meaning it is converted into a one-dimensional sequence of data.
  • The Position Attention Module (PAM) is applied, which allows the model to focus on important features in the sequence.

• Bottleneck:

  • The bottleneck processes the serialized feature maps to capture global dependencies.

• Expansion Phase (Decoder):

  • The decoder uses skip connections to integrate low-level feature information from the encoder with the high-level representations from the bottleneck.

Citing MPU-Net

If you use the code, please cite the article:

@article{zeqiu2023MP-UNet,
  author    = {Zeqiu Yu and Shuo Han},
  title     = {3D Medical Image Segmentation based on multi-scale MPU-Net},
  journal   = {arXiv preprint arXiv:2307.05799},
  year      = {2023},
  url       = {https://arxiv.org/abs/2307.05799}
}