Simultaneous Compression and De-Speckling of Medical Ultrasound Images

Compression of medical ultrasound images is necessary in order to meet storage and transmission requirements. Moreover, de-speckling is essential for diagnostic applications. We propose an algorithm for simultaneous compression and de-speckling of ultrasound images, based on optimization of quantization parameters applied on the coefficients of the coding transform. The optimization is applied in the rate-fidelity space, such that for a given target compression ratio, an optimal fidelity is obtained with respect to an a-priori known de-speckled image. The algorithm was tested on a set of fetal images. Results show that the reconstructed images better resemble the de-speckled images, compared to those obtained using standard quantization.