Interactive Segmentation for Diverse Gesture Types Without Context
This work addresses a problem for users of interactive segmentation tools by simplifying the input process, though it is incremental as it builds on existing methods with a new dataset and task formulation.
The paper tackles the limitation of existing interactive segmentation methods that require specifying gesture types and inclusion/exclusion, proposing a simplified task where users can mark images with any gesture type without such specifications. They introduce the first dataset with multiple gesture types and a new evaluation metric, analyzing various algorithms and observing promising performance while highlighting areas for improvement.
Interactive segmentation entails a human marking an image to guide how a model either creates or edits a segmentation. Our work addresses limitations of existing methods: they either only support one gesture type for marking an image (e.g., either clicks or scribbles) or require knowledge of the gesture type being employed, and require specifying whether marked regions should be included versus excluded in the final segmentation. We instead propose a simplified interactive segmentation task where a user only must mark an image, where the input can be of any gesture type without specifying the gesture type. We support this new task by introducing the first interactive segmentation dataset with multiple gesture types as well as a new evaluation metric capable of holistically evaluating interactive segmentation algorithms. We then analyze numerous interactive segmentation algorithms, including ones adapted for our novel task. While we observe promising performance overall, we also highlight areas for future improvement. To facilitate further extensions of this work, we publicly share our new dataset at https://github.com/joshmyersdean/dig.