Universal Cross-Domain Retrieval: Generalizing Across Classes and Domains

TL;DR

This paper introduces SnMpNet, a novel model for universal cross-domain retrieval that effectively generalizes to unseen classes and domains using semantic neighborhood and mix-up losses, validated on large datasets.

Contribution

The paper proposes SnMpNet with two new loss functions to improve generalization to unseen classes and domains in cross-domain retrieval tasks.

Findings

SnMpNet outperforms state-of-the-art methods on Sketchy Extended and DomainNet datasets.

Semantic Neighborhood loss enhances embedding quality for unseen classes.

Mix-up supervision improves retrieval accuracy across unseen domains.

Abstract

In this work, for the first time, we address the problem of universal cross-domain retrieval, where the test data can belong to classes or domains which are unseen during training. Due to dynamically increasing number of categories and practical constraint of training on every possible domain, which requires large amounts of data, generalizing to both unseen classes and domains is important. Towards that goal, we propose SnMpNet (Semantic Neighbourhood and Mixture Prediction Network), which incorporates two novel losses to account for the unseen classes and domains encountered during testing. Specifically, we introduce a novel Semantic Neighborhood loss to bridge the knowledge gap between seen and unseen classes and ensure that the latent space embedding of the unseen classes is semantically meaningful with respect to its neighboring classes. We also introduce a mix-up based supervision at image-level as well as semantic-level of the data for training with the Mixture Prediction loss, which helps in efficient retrieval when the query belongs to an unseen domain. These losses are incorporated on the SE-ResNet50 backbone to obtain SnMpNet. Extensive experiments on two large-scale datasets, Sketchy Extended and DomainNet, and thorough comparisons with state-of-the-art justify the effectiveness of the proposed model.