Compression of medical ultrasound images using wavelet transform and vector quantization

In this paper a new technique for lossy image compression of medical ultrasound images is proposed. The technique is based on the wavelet transform of the original image combined with the vector quantization (VQ) of high-energy subbands using LBG algorithm. The statistical analysis of wavelet coefficients in ultrasound images shows that the most of the image energy is concentrated in one of the detail subband either in vertical detail subband (most of the time) or in horizontal subband. The other two bands at each decomposition level contribute negligibly to the total image energy. This investigation allows us to achieve higher compression ratio by applying VQ only to the highest energy subband while discarding the other detail subbands at each level of decomposition. The technique was tested on a series of abdominal and uterus grayscale ultrasound images. The results indicate an average peak signal to noise (PSNR) ratio of 30.78 dB at an average compression ratio of 93:1. The proposed technique provides better diagnostic quality as determined by a new quality metric 'correlation-coefficient' as compared to the technique proposed by Xiang et al. even at much higher compression ratios.