Conformal Bounds on Full-Reference Image Quality for Imaging Inverse Problems
This work addresses a critical need in safety-critical applications like medical imaging by enabling reliable quality assessment without ground truth, though it is incremental as it builds on existing conformal prediction and sampling techniques.
The paper tackles the problem of estimating full-reference image quality metrics like PSNR or SSIM without access to the true image in imaging inverse problems, such as medical imaging, by developing a method that provides statistically guaranteed bounds on these metrics with a user-specified error probability.
In imaging inverse problems, we would like to know how close the recovered image is to the true image in terms of full-reference image quality (FRIQ) metrics like PSNR, SSIM, LPIPS, etc. This is especially important in safety-critical applications like medical imaging, where knowing that, say, the SSIM was poor could potentially avoid a costly misdiagnosis. But since we don't know the true image, computing FRIQ is non-trivial. In this work, we combine conformal prediction with approximate posterior sampling to construct bounds on FRIQ that are guaranteed to hold up to a user-specified error probability. We demonstrate our approach on image denoising and accelerated magnetic resonance imaging (MRI) problems. Code is available at https://github.com/jwen307/quality_uq.