Learning Equivariant Segmentation with Instance-Unique Querying

TL;DR

This paper introduces a new training framework for query-based instance segmentation models that enhances discriminative query learning by leveraging dataset-level uniqueness and transformation equivariance, resulting in improved performance.

Contribution

The proposed training algorithm improves query-based segmentation models by enforcing dataset-level uniqueness and transformation equivariance in query embeddings.

Findings

Significant AP improvements on COCO dataset (+1.6 - 3.2 AP)

Enhanced SOLOv2 performance on LVISv1 (+2.7 AP)

Applicable to multiple query-based models like CondInst, SOLOv2, SOTR, and Mask2Former.

Abstract

Prevalent state-of-the-art instance segmentation methods fall into a query-based scheme, in which instance masks are derived by querying the image feature using a set of instance-aware embeddings. In this work, we devise a new training framework that boosts query-based models through discriminative query embedding learning. It explores two essential properties, namely dataset-level uniqueness and transformation equivariance, of the relation between queries and instances. First, our algorithm uses the queries to retrieve the corresponding instances from the whole training dataset, instead of only searching within individual scenes. As querying instances across scenes is more challenging, the segmenters are forced to learn more discriminative queries for effective instance separation. Second, our algorithm encourages both image (instance) representations and queries to be equivariant against geometric transformations, leading to more robust, instance-query matching. On top of four famous, query-based models ($i.e.,$ CondInst, SOLOv2, SOTR, and Mask2Former), our training algorithm provides significant performance gains ($e.g.,$ +1.6 - 3.2 AP) on COCO dataset. In addition, our algorithm promotes the performance of SOLOv2 by 2.7 AP, on LVISv1 dataset.