Tuning Butterworth filter's parameters in SPECT reconstructions via kernel-based Bayesian optimization with a no-reference image evaluation metric

TL;DR

This paper introduces a kernel-based Bayesian optimization method utilizing a no-reference PIQUE score to tune Butterworth filter parameters in SPECT image reconstruction, enabling objective quality assessment without ground-truth images.

Contribution

It presents a novel application of PIQUE for SPECT image quality evaluation and integrates it with Bayesian optimization for parameter tuning, advancing objective image quality assessment methods.

Findings

Effective parameter tuning of Butterworth filters in SPECT reconstruction.

Demonstrated potential for objective, no-reference image quality assessment.

Improved clinical image quality evaluation without ground-truth images.

Abstract

In Single Photon Emission Computed Tomography (SPECT), the image reconstruction process involves many tunable parameters that have a significant impact on the quality of the resulting clinical images. Traditional image quality evaluation often relies on expert judgment and full-reference metrics such as MSE and SSIM. However, these approaches are limited by their subjectivity or the need for a ground-truth image. In this paper, we investigate the usage of a no-reference image quality assessment method tailored for SPECT imaging, employing the Perception-based Image QUality Evaluator (PIQUE) score. Precisely, we propose a novel application of PIQUE in evaluating SPECT images reconstructed via filtered backprojection using a parameter-dependent Butterworth filter. For the optimization of filter's parameters, we adopt a kernel-based Bayesian optimization framework grounded in reproducing kernel Hilbert space theory, highlighting the connections to recent greedy approximation techniques. Experimental results in a concrete clinical setting for SPECT imaging show the potential of this optimization approach for an objective and quantitative assessment of image quality, without requiring a reference image.