Training-free Linear Image Inversion via Flows

Ashwini Pokle, Matthew J. Muckley, Ricky T. Q. Chen, Brian Karrer

Keywords: inverse problems, linear image inversion, continuous normalizing flows, flow matching, diffusion models

TL;DR: We propose a training-free method for image inversion using pretrained flow models trained via flow matching.

Abstract:

Training-free linear inversion involves the use of a pretrained generative model and---through appropriate modifications to the generation process---solving inverse problems without any finetuning of the generative model.
While recent prior methods have explored the use of diffusion models, they still require the manual tuning of many hyperparameters for different inverse problems.
In this work, we propose a training-free method for image inversion using pretrained flow models, leveraging the simplicity and efficiency of Flow Matching models, using theoretically-justified weighting schemes and thereby significantly reducing the amount of manual tuning.
In particular, we draw inspiration from two main sources: adopting prior gradient correction methods to the flow regime, and a solver scheme based on conditional Optimal Transport paths.
As pretrained diffusion models are widely accessible, we also show how to practically adapt diffusion models for our method.
Empirically, our approach requires no problem-specific tuning across an extensive suite of noisy linear image inversion problems on high-dimensional datasets, ImageNet-64/128 and AFHQ-256, and we observe that our flow-based method for image inversion significantly improves upon closely-related diffusion-based linear inversion methods.

Primary Area: generative models

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 1983