Structured Query-Based Image Retrieval Using Scene Graphs
This addresses the challenge of more useful image retrieval for users needing complex queries, though it appears incremental as it builds on existing scene graph techniques.
The paper tackles the problem of image retrieval using structured queries that capture object interactions, such as 'woman rides motorcycle', and presents a method using scene graph embeddings to achieve high recall on low to medium frequency objects in the COCO-Stuff dataset, with a visual relationship-inspired loss boosting recall by 10% in the best case.
A structured query can capture the complexity of object interactions (e.g. 'woman rides motorcycle') unlike single objects (e.g. 'woman' or 'motorcycle'). Retrieval using structured queries therefore is much more useful than single object retrieval, but a much more challenging problem. In this paper we present a method which uses scene graph embeddings as the basis for an approach to image retrieval. We examine how visual relationships, derived from scene graphs, can be used as structured queries. The visual relationships are directed subgraphs of the scene graph with a subject and object as nodes connected by a predicate relationship. Notably, we are able to achieve high recall even on low to medium frequency objects found in the long-tailed COCO-Stuff dataset, and find that adding a visual relationship-inspired loss boosts our recall by 10% in the best case.