The effectiveness of the proposed MADR-Net architecture is evaluated with four publicly available datasets from different image modalities.
- Cardiac Acquisitions for Multi-Structure Ultrasound Segmentation (CAMUS) consist of a two-dimensional apical two-chamber and four-chamber view sequence.
- Dermoscopy images were acquired from the Medical Image Computing and Computer-Aided Intervention (MICCAI) conference hosted by the International Skin Imaging Collaboration (ISIC) in 2017 for analysis of skin lesions.
- FIB-SEM dataset consists of serial section transmission electron microscopy (SSTEM) images acquired from the hippocampus region of the brain.
- Brain MR images with manual ground truth of fluid-attenuated inversion recovery (FLAIR) abnormalities were acquired from The Cancer Imaging Archive (TCIA).
Progression of the validation dice score with respect to the number of epochs for Skin cancer dataset