Natural Language Guided Visual Relationship Detection
This addresses the challenge of detecting diverse object relationships in images for scene understanding, with incremental improvements in handling unseen relationships.
The paper tackles the problem of visual relationship detection by using natural language guidance to predict semantic connections between objects, achieving state-of-the-art results with recall improved from 76.42% to 89.79% on a zero-shot testing set.
Reasoning about the relationships between object pairs in images is a crucial task for holistic scene understanding. Most of the existing works treat this task as a pure visual classification task: each type of relationship or phrase is classified as a relation category based on the extracted visual features. However, each kind of relationships has a wide variety of object combination and each pair of objects has diverse interactions. Obtaining sufficient training samples for all possible relationship categories is difficult and expensive. In this work, we propose a natural language guided framework to tackle this problem. We propose to use a generic bi-directional recurrent neural network to predict the semantic connection between the participating objects in the relationship from the aspect of natural language. The proposed simple method achieves the state-of-the-art on the Visual Relationship Detection (VRD) and Visual Genome datasets, especially when predicting unseen relationships (e.g. recall improved from 76.42% to 89.79% on VRD zero-shot testing set).