Instance-Specific Feature Propagation for Referring Segmentation

TL;DR

This paper introduces a novel instance-specific feature propagation framework for referring segmentation that improves target identification and segmentation accuracy by modeling relationships among objects and integrating vision and language information.

Contribution

The paper proposes a new framework using Instance-Specific Features and a Feature Propagation Module to enhance referring segmentation performance.

Findings

Outperforms previous state-of-the-art on RefCOCO datasets.

Effectively models relationships among objects for better target localization.

Integrates vision and language information for improved segmentation accuracy.

Abstract

Referring segmentation aims to generate a segmentation mask for the target instance indicated by a natural language expression. There are typically two kinds of existing methods: one-stage methods that directly perform segmentation on the fused vision and language features; and two-stage methods that first utilize an instance segmentation model for instance proposal and then select one of these instances via matching them with language features. In this work, we propose a novel framework that simultaneously detects the target-of-interest via feature propagation and generates a fine-grained segmentation mask. In our framework, each instance is represented by an Instance-Specific Feature (ISF), and the target-of-referring is identified by exchanging information among all ISFs using our proposed Feature Propagation Module (FPM). Our instance-aware approach learns the relationship among all objects, which helps to better locate the target-of-interest than one-stage methods. Comparing to two-stage methods, our approach collaboratively and interactively utilizes both vision and language information for synchronous identification and segmentation. In the experimental tests, our method outperforms previous state-of-the-art methods on all three RefCOCO series datasets.