Densely Connected Parameter-Efficient Tuning for Referring Image Segmentation

TL;DR

DETRIS introduces a densely connected parameter-efficient tuning framework that enhances cross-modal feature interaction and adapts to misaligned encoders, significantly improving referring image segmentation performance with minimal parameter updates.

Contribution

The paper proposes a novel dense interconnection method for parameter-efficient tuning that effectively handles misaligned encoders in multimodal vision tasks.

Findings

Outperforms state-of-the-art methods on challenging benchmarks

Achieves 0.9% to 1.8% improvement with minimal parameter updates

Enhances cross-modal feature interaction and adaptation

Abstract

In the domain of computer vision, Parameter-Efficient Tuning (PET) is increasingly replacing the traditional paradigm of pre-training followed by full fine-tuning. PET is particularly favored for its effectiveness in large foundation models, as it streamlines transfer learning costs and optimizes hardware utilization. However, the current PET methods are mainly designed for single-modal optimization. While some pioneering studies have undertaken preliminary explorations, they still remain at the level of aligned encoders (e.g., CLIP) and lack exploration of misaligned encoders. These methods show sub-optimal performance with misaligned encoders, as they fail to effectively align the multimodal features during fine-tuning. In this paper, we introduce DETRIS, a parameter-efficient tuning framework designed to enhance low-rank visual feature propagation by establishing dense interconnections between each layer and all preceding layers, which enables effective cross-modal feature interaction and adaptation to misaligned encoders. We also suggest using text adapters to improve textual features. Our simple yet efficient approach greatly surpasses state-of-the-art methods with 0.9% to 1.8% backbone parameter updates, evaluated on challenging benchmarks. Our project is available at \url{https://github.com/jiaqihuang01/DETRIS}.