Attention UW-Net: A fully connected model for automatic segmentation and annotation of chest X-ray

Debojyoti Pal, Pailla Balakrishna Reddy, Sudipta Roy

Abstract: Highlights•Segmentation and annotation of chest X-ray plays a critical role in scientific healthcare research community.•Attention UW-Net has achieved high accuracy even from small data set.•This probabilistic map for automatic annotation helps to reduce the use of tedious and prone to error annotations.•A modified attention gate was implemented to further improve the prediction accuracy for segmentation of small lesions.•Modified attention vectors to reduces the number of redundant pixels predicted as a segmentation mask.