SaSR-Net: Source-Aware Semantic Representation Network for Enhancing Audio-Visual Question Answering

TL;DR

SaSR-Net is a novel model for Audio-Visual Question Answering that effectively captures and aligns multi-modal audio-visual information using source-aware tokens and attention mechanisms, leading to improved performance.

Contribution

The paper introduces SaSR-Net, a source-aware semantic network that enhances AVQA by using learnable tokens and attention for better multi-modal understanding.

Findings

Outperforms state-of-the-art AVQA methods on benchmark datasets.

Effectively captures audio-visual sources with source-wise tokens.

Utilizes spatial and temporal attention for improved scene understanding.

Abstract

Audio-Visual Question Answering (AVQA) is a challenging task that involves answering questions based on both auditory and visual information in videos. A significant challenge is interpreting complex multi-modal scenes, which include both visual objects and sound sources, and connecting them to the given question. In this paper, we introduce the Source-aware Semantic Representation Network (SaSR-Net), a novel model designed for AVQA. SaSR-Net utilizes source-wise learnable tokens to efficiently capture and align audio-visual elements with the corresponding question. It streamlines the fusion of audio and visual information using spatial and temporal attention mechanisms to identify answers in multi-modal scenes. Extensive experiments on the Music-AVQA and AVQA-Yang datasets show that SaSR-Net outperforms state-of-the-art AVQA methods.