A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondence
This work addresses the challenge of dense semantic matching for computer vision applications, offering a novel fusion approach that improves accuracy in tasks like instance swapping, though it is incremental in combining existing features.
The paper tackled the problem of zero-shot semantic correspondence across images by fusing Stable Diffusion (SD) and DINOv2 features, achieving significant performance gains over state-of-the-art methods on benchmark datasets like SPair-71k, PF-Pascal, and TSS.
Text-to-image diffusion models have made significant advances in generating and editing high-quality images. As a result, numerous approaches have explored the ability of diffusion model features to understand and process single images for downstream tasks, e.g., classification, semantic segmentation, and stylization. However, significantly less is known about what these features reveal across multiple, different images and objects. In this work, we exploit Stable Diffusion (SD) features for semantic and dense correspondence and discover that with simple post-processing, SD features can perform quantitatively similar to SOTA representations. Interestingly, the qualitative analysis reveals that SD features have very different properties compared to existing representation learning features, such as the recently released DINOv2: while DINOv2 provides sparse but accurate matches, SD features provide high-quality spatial information but sometimes inaccurate semantic matches. We demonstrate that a simple fusion of these two features works surprisingly well, and a zero-shot evaluation using nearest neighbors on these fused features provides a significant performance gain over state-of-the-art methods on benchmark datasets, e.g., SPair-71k, PF-Pascal, and TSS. We also show that these correspondences can enable interesting applications such as instance swapping in two images.