Generating Holistic 3D Scene Abstractions for Text-based Image Retrieval

TL;DR

This paper introduces a novel method for text-based image retrieval that infers 3D scene structures from textual descriptions to generate holistic object layouts, improving retrieval accuracy without requiring large annotated datasets.

Contribution

The paper proposes a framework that infers 3D scene layouts from text descriptions using physical relation models, enabling effective image retrieval based on holistic scene understanding.

Findings

Outperforms baselines based on object occurrence histograms.

Successfully infers 3D object layouts from text descriptions.

Validates approach on public indoor scene datasets.

Abstract

Spatial relationships between objects provide important information for text-based image retrieval. As users are more likely to describe a scene from a real world perspective, using 3D spatial relationships rather than 2D relationships that assume a particular viewing direction, one of the main challenges is to infer the 3D structure that bridges images with users' text descriptions. However, direct inference of 3D structure from images requires learning from large scale annotated data. Since interactions between objects can be reduced to a limited set of atomic spatial relations in 3D, we study the possibility of inferring 3D structure from a text description rather than an image, applying physical relation models to synthesize holistic 3D abstract object layouts satisfying the spatial constraints present in a textual description. We present a generic framework for retrieving images from a textual description of a scene by matching images with these generated abstract object layouts. Images are ranked by matching object detection outputs (bounding boxes) to 2D layout candidates (also represented by bounding boxes) which are obtained by projecting the 3D scenes with sampled camera directions. We validate our approach using public indoor scene datasets and show that our method outperforms baselines built upon object occurrence histograms and learned 2D pairwise relations.