Fig. 3
From: AttenUNeT X with iterative feedback mechanisms for robust deep learning skin lesion segmentation
Architecture of the attention module used in AttenUNeT X. The block performs spatial and channel attention before merging features in the decoder.
