Iterative Interaction Training for Segmentation Editing Networks
This work addresses the need for flexible editing tools in automatic segmentation for users in fields like medical imaging or computer vision, though it is incremental as it builds on existing interactive segmentation methods.
The paper tackles the problem of enabling interactive segmentation editing beyond binary cases by introducing a unique training strategy for CNNs that mimics realistic user interactions, resulting in substantial performance improvements with up to ten iterative interactions and achieving superior or on-par results compared to state-of-the-art methods.
Automatic segmentation has great potential to facilitate morphological measurements while simultaneously increasing efficiency. Nevertheless often users want to edit the segmentation to their own needs and will need different tools for this. There has been methods developed to edit segmentations of automatic methods based on the user input, primarily for binary segmentations. Here however, we present an unique training strategy for convolutional neural networks (CNNs) trained on top of an automatic method to enable interactive segmentation editing that is not limited to binary segmentation. By utilizing a robot-user during training, we closely mimic realistic use cases to achieve optimal editing performance. In addition, we show that an increase of the iterative interactions during the training process up to ten improves the segmentation editing performance substantially. Furthermore, we compare our segmentation editing CNN (interCNN) to state-of-the-art interactive segmentation algorithms and show a superior or on par performance.