Wavelet compression in medical terahertz pulsed imaging.

This paper concerns the robustness of discrete wavelet transform (DWT) compression in terahertz pulsed imaging (TPI). TPI datasets consist of terahertz time-domain series which are sampled at each 'pixel' of the image, leading to file sizes which are typically of the order of several megabytes (MB) per image. This makes efficient compression highly desirable for both transmission and storage. However, since the data may be required for diagnostic purposes it is essential that no relevant information is lost or artefacts introduced. We show that for a nylon step wedge the estimates of refractive index and absorption coefficients are not significantly altered when the terahertz data are reconstructed from only 20% of DWT coefficients.