Studies On Denoising And Enhancement Of Ultrasonic Medical Image

The ultrasonic medical imaging has become an important clinical diagnosis means because it has no damage on the human body, organ can display real-time, low cost, ease of use. However, Speckle noises are produced because of the principle of the ultrasonic medical imaging. The ultrasonic medical image detail information is not easy identified because of speckle noises. The detail information plays an important role in clinical diagnosis, so it has become an important research topic that the details of medical ultrasonic image are being preserved while speckle noises are being removed.In this paper, the characteristics of ultrasound medical images are analyzed, wavelet transform, morphology, pulse coupled neural networks, fuzzy algorithm is applied to medical ultrasonic image denoising and enhanced processing technology, to improve the medical ultrasonic image denoising and enhancing performance. The main research contents include the following aspects1. Study of the method of ultrasonic medical images Wavelet Threshold Denois-ingThe characteristics of de noising by soft threshold, semi-soft threshold and hard threshold were analyzed, according to the characteristics of medical ultrasonic image, a new de noising algorithm of medical ultrasonic image based on wavelet transform was proposed. And in view of the fact that VisuShrink threshold tends to "overkill" the detail coefficient and morphology can recognize the feature points of image, a de-noising algorithm of medical ultrasonic image based on morphology and wavelet transform is proposed. In the method, the high frequency coefficient which is less than the threshold is extracted by Hit-or-Miss Transform (HMT), and then the signal and noise are processed by improved threshold method. The experiment results show that the speckle noises are removed efficiently and the detail information is better reserved, farther more, the whole visual effect is improved.2. Study several methods of medical ultrasonic image denoising based on PCNN in the wavelet domainBased on the analysis of the properties of PCNN which can distinguish between signal and noise, Several methods of medical ultrasonic image denoising are proposed based on PCNN in the wavelet domain(1) A method of medical ultrasonic image de-noising based on PCNN in the Wavelet Domain (PCNN-WD) was proposed, via analyzing the characteristics of wavelet transform and PCNN. In the method, first, the wavelet coefficients were processed. After corresponding pretreatment, the wavelet coefficients were modified by PCNN in the wavelet domain so that the speckle noises can be removed, and the amplification coefficient of threshold and the step by which the coefficients of wavelet were modified can be automatic setting in this method. The experiment results show that the detail information and the image edge were being reserved when the speckle noises were being effectively removed by WT-PCNN.(2) Based on the analysis of the speckle noise and PCNN's properties, PCNN is introduced into wavelet domain, by combining the thought of the soft-threshold de-noising, Soft-Threshold de-noising method of medical ultrasonic image based on PCNN(ST-PCNN) is proposed. The advantage of ST-PCNN is that PCNN recognizes the coefficients of high frequency in wavelet domain are realized, and then the wavelet coefficients are processed by corresponding methods. ST-PCNN improves the disadvantage that PCNN can not accurately determine the position of speckle noise and the fixed threshold makes some high-frequency signals loss, and better reserves the wavelet coefficients of high frequency signal which are lower than the fixed threshold. On this basis, fuzzy algorithm is applied in the model of PCNN, method of medical ultrasonic image de-noising based on Fuzzy PCNN in the Wavelet Domain (F-PCNN-WD) is proposed.The proposed method make use of fuzzy algorithm to remove the wavelet coefficients of speckle noise which are greater than the ignition threshold value of PCNN,so the speckle noise can be better removed The experimental results show that ST-PCNN and F-PCNN-WD can not only remove the noise but also reserve the detail information and the image edge.3. Study a fusion method of medical ultrasonic image denoising based on wiener filter and wavelet transformthe feature of wiener filter and Multi-scale Speckle Suppression by Nonlinear Thresholding based on Adaptive presholding (MSSNT-A) are analyzed. A method of medical ultrasonic image denoising based on wiener filter and MSSNT-A is proposed. In this method,at first. the noise image was respectively denoised by wiener filter and MSSNT-A; then the edge of image which was denoised by wiener filter is extracted; At the end, the edge and the image that was denoised by MSSNT-A were fused together, the final denoised image was obtained. The experiment results show that,compared with wiener filter and MSSNT-A,this method can better retention the edge and details of the image.4. Study the method of medical ultrasonic image enhancement based on adaptive low pass filterAccording to the "degenerate" phenomenon of full frame histogram and the blindness enhanced images of local area histogram,a method of medical ultrasonic image enhancement based on adaptive low pass filter is proposed. In this method,at first, Logarithmic transform was carried out to the medical ultrasound image. Multiplicative noises were transformed into additive ones. The high and low frequency parts were departed with the help of low pass filter. Then the low frequency component was processed with adaptive local area histogram equalization algorithm and the high component was weighted.In the end, the two components were reunited to get the enhanced logarithmic image, the exponential transform were processed.