MultiResUNet model also uses Transposed Convolutions in the encoder by default instead of UpSampling. Internal structure for the MultiResUNet block is shown below [4]. As it can be seen, the MultiRes ...

using the Unet model to segment images in order to find which lanes are drivable for a car - UNet_Drivable_area_Segmentation/README.md at main · CoNfIg7952/UNet_Drivable_area_Segmentation. ... Encoder ...