Evaluating Contrastive Models for Instance-based Image Retrieval
This provides a practical solution for building image retrieval systems without labeled data, though it is incremental as it evaluates existing methods rather than proposing new ones.
The paper evaluated contrastive learning models for instance-based image retrieval, finding they perform on-par with or outperform supervised ImageNet-trained baselines without requiring explicit supervision.
In this work, we evaluate contrastive models for the task of image retrieval. We hypothesise that models that are learned to encode semantic similarity among instances via discriminative learning should perform well on the task of image retrieval, where relevancy is defined in terms of instances of the same object. Through our extensive evaluation, we find that representations from models trained using contrastive methods perform on-par with (and outperforms) a pre-trained supervised baseline trained on the ImageNet labels in retrieval tasks under various configurations. This is remarkable given that the contrastive models require no explicit supervision. Thus, we conclude that these models can be used to bootstrap base models to build more robust image retrieval engines.