R-VQA: Learning Visual Relation Facts with Semantic Attention for Visual Question Answering

TL;DR

This paper introduces R-VQA, a framework that leverages visual relation facts with semantic attention to improve visual question answering, achieving state-of-the-art results by integrating semantic knowledge and visual relations.

Contribution

The paper proposes a novel R-VQA framework that learns and utilizes visual relation facts with semantic attention, enhancing VQA performance beyond existing methods.

Findings

Achieves state-of-the-art results on benchmark datasets.

Demonstrates the effectiveness of visual relation facts in VQA.

Shows benefits of semantic attention in integrating knowledge.

Abstract

Recently, Visual Question Answering (VQA) has emerged as one of the most significant tasks in multimodal learning as it requires understanding both visual and textual modalities. Existing methods mainly rely on extracting image and question features to learn their joint feature embedding via multimodal fusion or attention mechanism. Some recent studies utilize external VQA-independent models to detect candidate entities or attributes in images, which serve as semantic knowledge complementary to the VQA task. However, these candidate entities or attributes might be unrelated to the VQA task and have limited semantic capacities. To better utilize semantic knowledge in images, we propose a novel framework to learn visual relation facts for VQA. Specifically, we build up a Relation-VQA (R-VQA) dataset based on the Visual Genome dataset via a semantic similarity module, in which each data consists of an image, a corresponding question, a correct answer and a supporting relation fact. A well-defined relation detector is then adopted to predict visual question-related relation facts. We further propose a multi-step attention model composed of visual attention and semantic attention sequentially to extract related visual knowledge and semantic knowledge. We conduct comprehensive experiments on the two benchmark datasets, demonstrating that our model achieves state-of-the-art performance and verifying the benefit of considering visual relation facts.