LiveSketch: Query Perturbations for Guided Sketch-Based Visual Search

LiveSketch is a novel algorithm for searching large image collections using hand-sketched queries. LiveSketch tackles the inherent ambiguity of sketch search by creating visual suggestions that augment the query as it is drawn, making query specification an iterative rather than one-shot process that helps disambiguate users' search intent. Our technical contributions are: a triplet convnet architecture that incorporates an RNN based variational autoencoder to search for images using vector (stroke-based) queries; real-time clustering to identify likely search intents (and so, targets within the search embedding); and the use of backpropagation from those targets to perturb the input stroke sequence, so suggesting alterations to the query in order to guide the search. We show improvements in accuracy and time-to-task over contemporary baselines using a 67M image corpus.

1. Introduction

Determining user intent from a visual search query remains an open challenge, particularly in sketch based image retrieval (SBIR) over millions of images where a sketched shape can yield plausible yet unexpected matches. For example, a user’s sketch of a dog might return a map of the United States that ostensibly resembles the shape (structure) drawn, but is not relevant. Free-hand sketches are often incomplete and ambiguous descriptions of desired image content [8]. This limits the ability of sketch to communicate search intent, particularly over large image datasets.