GRES: Generalized Referring Expression Segmentation

TL;DR

This paper introduces GRES, a new benchmark and dataset for generalized referring expression segmentation that includes multi-target, no-target, and single-target expressions, and proposes a novel method ReLA that improves performance.

Contribution

The paper presents the first large-scale GRES dataset gRefCOCO and a new baseline ReLA that models complex relationships for improved segmentation.

Findings

ReLA achieves state-of-the-art results on GRES and classic RES tasks.

GRES dataset includes multi-target, no-target, and single-target expressions.

Complex relationship modeling is a key challenge in GRES.

Abstract

Referring Expression Segmentation (RES) aims to generate a segmentation mask for the object described by a given language expression. Existing classic RES datasets and methods commonly support single-target expressions only, i.e., one expression refers to one target object. Multi-target and no-target expressions are not considered. This limits the usage of RES in practice. In this paper, we introduce a new benchmark called Generalized Referring Expression Segmentation (GRES), which extends the classic RES to allow expressions to refer to an arbitrary number of target objects. Towards this, we construct the first large-scale GRES dataset called gRefCOCO that contains multi-target, no-target, and single-target expressions. GRES and gRefCOCO are designed to be well-compatible with RES, facilitating extensive experiments to study the performance gap of the existing RES methods on the GRES task. In the experimental study, we find that one of the big challenges of GRES is complex relationship modeling. Based on this, we propose a region-based GRES baseline ReLA that adaptively divides the image into regions with sub-instance clues, and explicitly models the region-region and region-language dependencies. The proposed approach ReLA achieves new state-of-the-art performance on the both newly proposed GRES and classic RES tasks. The proposed gRefCOCO dataset and method are available at https://henghuiding.github.io/GRES.