UniST: Towards Unifying Saliency Transformer for Video Saliency Prediction and Detection

TL;DR

UniST introduces a unified transformer framework that effectively models both video saliency prediction and salient object detection, achieving superior results across multiple benchmarks by leveraging multi-scale spatio-temporal features.

Contribution

This work is the first to design a transformer-based model that jointly addresses both video saliency prediction and salient object detection tasks.

Findings

Achieves superior performance on seven benchmarks.

Outperforms existing state-of-the-art methods.

Effectively models both tasks with a unified framework.

Abstract

Video saliency prediction and detection are thriving research domains that enable computers to simulate the distribution of visual attention akin to how humans perceiving dynamic scenes. While many approaches have crafted task-specific training paradigms for either video saliency prediction or video salient object detection tasks, few attention has been devoted to devising a generalized saliency modeling framework that seamlessly bridges both these distinct tasks. In this study, we introduce the Unified Saliency Transformer (UniST) framework, which comprehensively utilizes the essential attributes of video saliency prediction and video salient object detection. In addition to extracting representations of frame sequences, a saliency-aware transformer is designed to learn the spatio-temporal representations at progressively increased resolutions, while incorporating effective cross-scale saliency information to produce a robust representation. Furthermore, a task-specific decoder is proposed to perform the final prediction for each task. To the best of our knowledge, this is the first work that explores designing a transformer structure for both saliency modeling tasks. Convincible experiments demonstrate that the proposed UniST achieves superior performance across seven challenging benchmarks for two tasks, and significantly outperforms the other state-of-the-art methods.