IDEA: An Invariant Perspective for Efficient Domain Adaptive Image Retrieval

In this paper, we study the problem of unsupervised domain adaptive retrieval, which transfers retrieval models from a label-rich source domain to a label-scarce target domain. Although there exist numerous approaches that incorporate transfer learning techniques into deep hashing frameworks, they often overlook the crucial invariance needed for adequate alignment between these two domains. Even worse, these methods fail to distinguish between causal and non-causal effects embedded in images, making cross-domain retrieval ineffective. To address these challenges, we propose an Invariance-acquired Domain Adaptive Hashing (IDEA) model. Our IDEA first decomposes each image into a causal feature representing label information and a non-causal feature indicating domain information. We then generate discriminative hash codes using causal features with consistency learning on both source and target domains. More importantly, we employ a generative model for synthetic samples to simulate the intervention of various non-causal effects, thereby minimizing their impact on hash codes for domain invariance. Comprehensive experiments conducted on benchmark datasets confirm the superior performance of our proposed IDEA compared to a variety of competitive baselines.