Taguchi's experimental design method on improvement of medical image quality

Abstract

Purpose - The aim of this article is to show how Taguchi methods can be applied to health care to improve the quality of medical images. Quality is often integrated with the performance and parameters of the design of medical applications. Many imaging methods can be designed by setting the correct combination of parameters and estimating the contribution of individual quality influencing factors by means of incorporating parameter design and orthogonal arrays. The performance of any imaging equipment can be measured by signal-to-noise ratio. This inherent index can give a sense of how close the performance is to the ideal. Design/methodology/approach - Data were collected from a database of 82 diagnostic thoracic computed tomography (CT) scans. Signal-to-noise ratios (S/N) were calculated. Findings - Given the S/N's, the best CT level was found to be level 4. Originality/value - To reduce bias resulting from the observer's readings, robust equipments should be designed incorporating Taguchi's experimental design. Further work is needed to establish imaging protocols and new hardware design.