Attention U-Net
Attention U-Net: Learning Where to Look for the Pancreas
We propose a novel attention gate (AG) model for medical imaging that automatically learns to focus on target structures of varying shapes and sizes. Models trained with AGs implicitly learn to suppress irrelevant regions in an input image while highlighting salient features useful for a specific task. This enables us to eliminate the necessity of using explicit external tissue/organ localisation modules of cascaded convolutional neural networks (CNNs). AGs can be easily integrated into standard CNN architectures such as the U-Net model with minimal computational overhead while increasing the model sensitivity and prediction accuracy. The proposed Attention U-Net architecture is evaluated on two large CT abdominal datasets for multi-class image segmentation. Experimental results show that AGs consistently improve the prediction performance of U-Net across different datasets and training sizes while preserving computational efficiency. The source code for the proposed architecture is publicly available.
Implementations
Pytorch Implementation of U-Net, R2U-Net, Attention U-Net, Attention R2U-Net
Evaluation
we just test the models with ISIC 2018 dataset. The dataset was split into three subsets, training set, validation set, and test set, which the proportion is 70%, 10% and 20% of the whole dataset, respectively. The entire dataset contains 2594 images where 1815 images were used for training, 259 for validation and 520 for testing models.
Model: https://github.com/LeeJunHyun/Image_Segmentation/blob/master/network.py
Github: https://github.com/LeeJunHyun/Image_Segmentation#attention-u-net
