Boosting Salient Object Detection with Knowledge Distillated from Large Foundation Models

TL;DR

This paper introduces a novel approach for salient object detection that leverages large foundation models for efficient pseudo-label generation, resulting in a new dataset and a method that outperforms existing techniques.

Contribution

We propose a weakly supervised method using large models and textual prompts to generate high-quality pseudo-labels, along with a new, larger dataset BDS-TR and an edge decoder for improved detection.

Findings

Our method outperforms state-of-the-art SOD approaches on five benchmarks.

The BDS-TR dataset is larger and more diverse than previous datasets.

The edge decoder enhances boundary accuracy in salient object detection.

Abstract

Salient Object Detection (SOD) aims to identify and segment prominent regions within a scene. Traditional models rely on manually annotated pseudo labels with precise pixel-level accuracy, which is time-consuming. We developed a low-cost, high-precision annotation method by leveraging large foundation models to address the challenges. Specifically, we use a weakly supervised approach to guide large models in generating pseudo-labels through textual prompts. Since large models do not effectively focus on the salient regions of images, we manually annotate a subset of text to fine-tune the model. Based on this approach, which enables precise and rapid generation of pseudo-labels, we introduce a new dataset, BDS-TR. Compared to the previous DUTS-TR dataset, BDS-TR is more prominent in scale and encompasses a wider variety of categories and scenes. This expansion will enhance our model's applicability across a broader range of scenarios and provide a more comprehensive foundational dataset for future SOD research. Additionally, we present an edge decoder based on dynamic upsampling, which focuses on object edges while gradually recovering image feature resolution. Comprehensive experiments on five benchmark datasets demonstrate that our method significantly outperforms state-of-the-art approaches and also surpasses several existing fully-supervised SOD methods. The code and results will be made available.