Wavelets improve medical imaging

The conventional approach to processing imaging data from medical scanners is to apply a computer algorithm, a Fourier transform, which converts the raw signals into a format that can be displayed as an image on the computer screen to reveal brain damage, detect tumors and look for defects in the heart and for countless other diagnostic reasons. The Fourier transform is a powerful tool but Satya Singh and Shabana Urooj of the Gautam Buddha University, in Uttar Pradesh, India, suggest that an alternative signal processing system based on an analysis of "wavelets" could improve medical imaging significantly. They point out that a wavelet transform applied to the raw imaging data would boost the signal to noise ratio for the data by ignoring many of the artifacts generated by interfering electrical signals and waves present in the body and picked by the scanner.

The team reports that the same wavelet transform approach could have applications not only in MRI but in simplifying electrocardiography, de-noising biomedical images and in facilitating and functional neuroimaging, including positron emission tomography and functional MRI (fMRI). The latter requiring data processing in real-time so that live images can be recorded and studied. In addition, the wavelet transform produces a much smaller digital file than conventional processing, which means faster data transfer and lower storage requirements per scan.