Fig. 10: Basic workflow of the Mask R-CNN algorithm.
From: A deep learned nanowire segmentation model using synthetic data augmentation
The input image is processed by different network components to extract high-level features, region of interests for prediction of bounding box, classes and masks.
