Visual masking and the design of magnetic resonance image acquisition

This paper presents novel data acquisition schemes for improving the quality of magnetic resonance images (MRI). These approaches determine the number of times that k-space (frequency domain) samples have to be acquired and averaged to form an MRI image, by taking into account the characteristics of the human visual system and the location in k-space of the data samples. They lead to dramatic reductions in imaging time (by a factor of 8 to 10) with no reduction in image quality. Alternatively, they produce higher quality imagery using the same amount of imaging time as conventional approaches.