Optimizing detector size in X-ray imaging

Image data are noisy samples acquired at discrete positions. Signal processing explicitly or implicitly involves some kind of interpolation to match a desired point spread function (PSF), to match a certain spatial resolution or to match certain contrast detection criteria. By optimization of a resolution-, noise- and dose-dependent figure of merit it was recently shown that a given spatial resolution can be best achieved using detector elements far smaller than the desired resolution element. A potential dose reduction in the order of 50% was predicted when using detector elements of about half the size of the desired spatial resolution. Instead of prescribing spatial resolution we here investigate how a square wave pattern of given detail size can be best imaged in terms of contrast, noise and dose. A contrast-to-noise quality factor Q is maximized to find the optimal detector element size g. Our findings are a potential of 70% dose reduction when using optimized flat panel detector pixel size g. This optimum lies in the order of 30% of the object detail size.