Interpolating between Images with Diffusion Models
This work addresses a missing feature in image generation pipelines, potentially expanding creative applications, though it is incremental as it builds on existing latent diffusion models.
The paper tackles the problem of interpolating between two input images using diffusion models, achieving convincing results across diverse subject poses, styles, and content, while noting that standard metrics like FID are insufficient for evaluating interpolation quality.
One little-explored frontier of image generation and editing is the task of interpolating between two input images, a feature missing from all currently deployed image generation pipelines. We argue that such a feature can expand the creative applications of such models, and propose a method for zero-shot interpolation using latent diffusion models. We apply interpolation in the latent space at a sequence of decreasing noise levels, then perform denoising conditioned on interpolated text embeddings derived from textual inversion and (optionally) subject poses. For greater consistency, or to specify additional criteria, we can generate several candidates and use CLIP to select the highest quality image. We obtain convincing interpolations across diverse subject poses, image styles, and image content, and show that standard quantitative metrics such as FID are insufficient to measure the quality of an interpolation. Code and data are available at https://clintonjwang.github.io/interpolation.