Semantic-Aware Knowledge Preservation for Zero-Shot Sketch-Based Image Retrieval
This work addresses a challenging setting in sketch-based image retrieval for applications requiring generalization to unseen categories, but it is incremental as it builds on existing domain adaptation frameworks.
The paper tackles the problem of zero-shot sketch-based image retrieval by proposing a method to preserve knowledge from pre-trained models, achieving superior performance on TU-Berlin and Sketchy datasets.
Sketch-based image retrieval (SBIR) is widely recognized as an important vision problem which implies a wide range of real-world applications. Recently, research interests arise in solving this problem under the more realistic and challenging setting of zero-shot learning. In this paper, we investigate this problem from the viewpoint of domain adaptation which we show is critical in improving feature embedding in the zero-shot scenario. Based on a framework which starts with a pre-trained model on ImageNet and fine-tunes it on the training set of SBIR benchmark, we advocate the importance of preserving previously acquired knowledge, e.g., the rich discriminative features learned from ImageNet, to improve the model's transfer ability. For this purpose, we design an approach named Semantic-Aware Knowledge prEservation (SAKE), which fine-tunes the pre-trained model in an economical way and leverages semantic information, e.g., inter-class relationship, to achieve the goal of knowledge preservation. Zero-shot experiments on two extended SBIR datasets, TU-Berlin and Sketchy, verify the superior performance of our approach. Extensive diagnostic experiments validate that knowledge preserved benefits SBIR in zero-shot settings, as a large fraction of the performance gain is from the more properly structured feature embedding for photo images. Code is available at: https://github.com/qliu24/SAKE.