ContrastMask: Contrastive Learning to Segment Every Thing
This addresses the problem of reducing annotation burden for instance segmentation in computer vision, though it is incremental as it builds on existing contrastive learning frameworks.
The paper tackles partially-supervised instance segmentation by proposing ContrastMask, a method that uses contrastive learning on both seen and unseen categories to build a class-agnostic mask segmentation model, achieving state-of-the-art results on the COCO dataset.
Partially-supervised instance segmentation is a task which requests segmenting objects from novel unseen categories via learning on limited seen categories with annotated masks thus eliminating demands of heavy annotation burden. The key to addressing this task is to build an effective class-agnostic mask segmentation model. Unlike previous methods that learn such models only on seen categories, in this paper, we propose a new method, named ContrastMask, which learns a mask segmentation model on both seen and unseen categories under a unified pixel-level contrastive learning framework. In this framework, annotated masks of seen categories and pseudo masks of unseen categories serve as a prior for contrastive learning, where features from the mask regions (foreground) are pulled together, and are contrasted against those from the background, and vice versa. Through this framework, feature discrimination between foreground and background is largely improved, facilitating learning of the class-agnostic mask segmentation model. Exhaustive experiments on the COCO dataset demonstrate the superiority of our method, which outperforms previous state-of-the-arts.