Metric-Guided Conformal Bounds for Probabilistic Image Reconstruction

Matt Y. Cheung, Tucker James Netherton, Laurence Edward Court, Ashok Veeraraghavan, Guha Balakrishnan

Keywords: Image Reconstruction · Conformal Prediction · Sparse-view CT · Deep Generative Models

Abstract:

Modern deep learning reconstruction algorithms generate
impressively realistic scans from sparse inputs, but can often produce
significant inaccuracies. This makes it difficult to provide statistically
guaranteed claims about the true state of a subject from scans reconstructed
by these algorithms. In this study, we propose a framework
for computing provably valid prediction bounds on claims derived from
probabilistic black-box image reconstruction algorithms. The key insights
are to represent reconstructed scans with a derived clinical metric of
interest and to calibrate bounds on the ground truth metric with conformal
prediction (CP) using a prior calibration dataset. These bounds
convey interpretable feedback about the subject’s state, and can also
be used to retrieve nearest-neighbor reconstructed scans for visual inspection.
We demonstrate the utility of this framework on sparse-view
computed tomography (CT) for fat mass quantification and radiotherapy
planning tasks. Results show that our framework produces visual
bounds with better semantical interpretation than conventional pixelbased
bounding approaches and captures important spatial correlations.
Furthermore, we can flag dangerous outlier reconstructions that look
plausible but have statistically unlikely metric values. Code available at:
https://github.com/matthewyccheung/conformal-metric

Submission Number: 16