The encoder progressively compresses spatial features, while the decoder restores resolution. Lateral skip connections pass fine-grained information to the decoder, making the design useful for segmentation and diffusion-model denoising.
A U-Net is an encoder-decoder architecture with skip connections between matching resolution levels.
The encoder progressively compresses spatial features, while the decoder restores resolution. Lateral skip connections pass fine-grained information to the decoder, making the design useful for segmentation and diffusion-model denoising.