UCP-Net: Unstructured Contour Points for Instance Segmentation
This addresses the problem of reducing user effort in producing segmentation masks for applications like image editing or medical imaging, though it appears incremental as it builds on existing interactive segmentation methods.
The paper tackles interactive segmentation by proposing a novel approach using unconstrained contour clicks for initial segmentation and refinement, achieving accurate masks (IoU > 85%) with fewer user interactions than state-of-the-art methods on datasets like COCO MVal, SBD, and Berkeley.
The goal of interactive segmentation is to assist users in producing segmentation masks as fast and as accurately as possible. Interactions have to be simple and intuitive and the number of interactions required to produce a satisfactory segmentation mask should be as low as possible. In this paper, we propose a novel approach to interactive segmentation based on unconstrained contour clicks for initial segmentation and segmentation refinement. Our method is class-agnostic and produces accurate segmentation masks (IoU > 85%) for a lower number of user interactions than state-of-the-art methods on popular segmentation datasets (COCO MVal, SBD and Berkeley).