Single image signal-to-noise ratio estimation for magnetic resonance images

A novel technique to quantify the signal-to-noise ratio (SNR) of magnetic resonance images is developed. The image SNR is quantified by estimating the amplitude of the signal spectrum using the autocorrelation function of just one single magnetic resonance image. To test the performance of the quantification, SNR measurement data are fitted to theoretically expected curves. It is shown that the technique can be implemented in a highly efficient way for the magnetic resonance imaging system.