How to Handle Sketch-Abstraction in Sketch-Based Image Retrieval?

TL;DR

This paper introduces an abstraction-aware sketch-based image retrieval framework that models sketch abstraction holistically, leveraging pre-trained StyleGAN features and a novel loss to improve retrieval across varied abstraction levels.

Contribution

It proposes a new framework that models sketch abstraction at multiple levels using feature and retrieval granularity designs, incorporating StyleGAN embeddings and a differentiable loss.

Findings

Outperforms existing state-of-the-art SBIR methods.

Effective in challenging scenarios like early retrieval and style-invariant matching.

Demonstrates robustness across different abstraction levels.

Abstract

In this paper, we propose a novel abstraction-aware sketch-based image retrieval framework capable of handling sketch abstraction at varied levels. Prior works had mainly focused on tackling sub-factors such as drawing style and order, we instead attempt to model abstraction as a whole, and propose feature-level and retrieval granularity-level designs so that the system builds into its DNA the necessary means to interpret abstraction. On learning abstraction-aware features, we for the first-time harness the rich semantic embedding of pre-trained StyleGAN model, together with a novel abstraction-level mapper that deciphers the level of abstraction and dynamically selects appropriate dimensions in the feature matrix correspondingly, to construct a feature matrix embedding that can be freely traversed to accommodate different levels of abstraction. For granularity-level abstraction understanding, we dictate that the retrieval model should not treat all abstraction-levels equally and introduce a differentiable surrogate Acc.@q loss to inject that understanding into the system. Different to the gold-standard triplet loss, our Acc.@q loss uniquely allows a sketch to narrow/broaden its focus in terms of how stringent the evaluation should be - the more abstract a sketch, the less stringent (higher q). Extensive experiments depict our method to outperform existing state-of-the-arts in standard SBIR tasks along with challenging scenarios like early retrieval, forensic sketch-photo matching, and style-invariant retrieval.