The Methods Of Tissue Damage Assessment Based On B-mode Ultrasound Imaging During HIFU Treatment

High intensity focused ultrasound(HIFU)technology is a non-invasive treat-ment technique in the treatment of tumors.Because it is non-invasive,it is of universality and practicality in clinical treatment of tumors.The high intensity focused ultrasound can be used to treat tumor in a noninvasive or minimally in-vasive manner,mainly based on its high strength and accuracy.Ultrasound can accurately focus energy in the target area of the tumor,causing instantaneous high temperature.Thus,it can cause irreversible coagulation necrosis or degener-ation in target area tissues.In the treatment of HIFU,tissue damage degree is an important index to evaluate the efficacy of this technique.In this paper,HIFU irradiated fresh pork tissue in vitro,which was used to simulate the process of tumor treatment.Ultrasound images of pork tissues before and after treatment were obtained through B-mode ultrasound diagnostic set,so as to study the change of image characteristics and evaluate the tissue damage caused by HIFU.The main tasks are as follows.Firstly,based on a large number of experiments,the obtained image data was studied and analyzed in detail,including the screening and pretreatment of image data.Based on image characteristics,the general rule of HIFU treatment on the degree of biological tissue damage was verified.Secondly,the multiple image characteristics of the image after pretreatment were extracted,including the mean of gray,variance and Hu moment invariant.In this paper,support vector machine was introduced to determine the classification of tissue degeneration qualitatively after treatment.At the same time,the wavelet transform and characteristic parameters were combined to introduce mean of the wavelet coefficients,variance of the wavelet coefficients and Hu moment of the wavelet coefficients.The results show that there is little difference between Hu moment of the wavelet coefficients and variance of the wavelet coefficients in the total identification rate.The identification effect of these two parameters is better in six parameters.Thirdly,three levels of tissue damage rating model was established to provide theory support of data for secondary HIFU irradiation.Different HIFU doses can cause different damage areas to target area tissue.According to the actual damage area,the areas were divided into three grades:no solidification heat damage,basic anastomosis between the coagulation damage area and focal area,and excessive damage.Fourthly,the gray mean and the mean of the wavelet coefficients were ex-tracted,and the k-means clustering method was introduced.Thus,the tissue dam-age was automatically identified as three levels according to the characteristic parameters.In addition,the clustering effects based on single parameter and dou-ble parameters were compared.The results show that the clustering method based on double parameters is better than single parameter clustering method,and it can judge the damage grade of the organization more accurately.Fifthly,the three-classification decision method was expanded.And the three classification decisions were made according to the distribution of parameter val-ues,rather than the area of damage measured by human measurement.Firstly,the gray mean and the mean value of the wavelet coefficients were extracted,then SVD processing was carried out.Finally,an AP clustering algorithm was used to achieve a more objective division of organizational damage level.Stand in the perspective of image processing in this paper,combined with machine learning methods,the evaluation of tissue damage degree after HIFU was achieved,which provided a new direction for future research.Meanwhile,there also was a certain practical significance on improving the HIFU therapeutic effect.