Unconstrained Foreground Object Search
This addresses a limitation for users in image editing by enabling more flexible object retrieval, though it is incremental as it builds on existing retrieval methods.
The paper tackles the problem of searching for foreground objects without semantic class constraints in image editing, proposing an unconstrained foreground object search solution that encodes background images in the same latent space as candidates, with experiments showing advantages over baselines.
Many people search for foreground objects to use when editing images. While existing methods can retrieve candidates to aid in this, they are constrained to returning objects that belong to a pre-specified semantic class. We instead propose a novel problem of unconstrained foreground object (UFO) search and introduce a solution that supports efficient search by encoding the background image in the same latent space as the candidate foreground objects. A key contribution of our work is a cost-free, scalable approach for creating a large-scale training dataset with a variety of foreground objects of differing semantic categories per image location. Quantitative and human-perception experiments with two diverse datasets demonstrate the advantage of our UFO search solution over related baselines.