Medical clinical ultrasound image segmentation based on modified wavelet transformation method

Ultrasound medical image has drawn extensive attentions for clinical monitoring and pharmaceutics determination, and the quality of image plays essential role for accurate information interpretation. In order to improve the contrast of the ultrasound medical image, algorithm based on modified wavelet transformation method has been established and optimized in this article. The modules maximum criterion is proposed for segmenting edges from the blurring background, though many pseudo-edges appear. Definition on the continuous of the edge according to the amount as P of the connected pixels implies that the higher coefficient P has induced the preservation of longer edges, and shorter segmented edges with fewer pixels have been erased. Entropy calculation indicates that the lower P has arisen out more decrease entropy deviation, and optimal threshold of P coefficient is defined at the 10% entropy deviation from the original ultrasound image. Consequently, the ultrasound image has been segmented out with more essential edges by the combined criterions of modules maximum, edge continuous and optimal entropy. The obtained algorithm can be integrated into the clinical evaluation software for ultrasound image interpretation to enhance the diagnostics and pharmacology accuracy.