Research On Automatic Analysis Techniques Of Intima-Media Of The Carotid Artery Based On Ultrasonic Radiofrequencv Signal

Carotid intima-media thickness(IMT)is an important indicator of early detection of atherosclerosis.Obtaining IMT is helpful for doctors to predict atherosclerosis.It is important for quantitative monitoring and analysis of carotid intima-media.In clinical,ultrasound imaging is often used to obtain the structure of the carotid artery,and the IMT is measured by the doctor manually or by boundary depiction.Due to the time consuming characteristic of the tedious manual calibration mode,which also leads to subjective error,it is necessary that the computer aided diagnosis,combined with image processing technology,be used to achieve a stable and accurate automatic segmentation of intima-media complex.In this thesis,an algorithm for the enhancement and extraction of the intima-media complex based on the analysis of radio frequency signal(RF)is presented.The algorithm,with original ultrasonic data radio frequency(RF)as the foundation,can be divided into three steps:First,the region of interest(ROI)is automatically determined in the first frame of the video sequence,and then the boundaries of the intima-media membranes are extracted by the delay estimation algorithm,which estimate the time delay between the adjacent scan lines by the RF data at the starting seed point of the automatic calibration;Second,on the subsequent frames,an improved normalized correlation algorithm is used to track the target points determined on the first frame;Finally,because of the ambiguous and discontinuous characteristics of the intima membrane image,an intima membrane enhancement algorithm in RF domain is proposed.Firstly,the inner membrane boundary is refined by using ant colony optimization algorithm,and then the RF signal is enhanced by using the delay-sum algorithm of RF signal scanning line on each frame.In the non-inner membrane boundary region,the filtering de-noising is carried out and eventually achieve the purpose of enhancing the inner membrane.The clinical experimental results show that the Pearson correlation coefficient between the IMT results of the proposed algorithm and the manual measurements of ultrasound clinician is 0.98,the autocorrelation coefficient is up to 0.99.The IMT measurement algorithm has high feasibility and good stability for clinical quantitative inspection and monitoring of carotid IMT measurement.The tracking algorithm proposed in this paper can be used for the dynamic detection of IMT,and the tracking error is kept within 0.15mm.Meanwhile,the experimental results show that the proposed RF domain internal membrane enhancement algorithm has a better performance.In addition,this set of algorithms can dynamically monitor the changes of intima-media in carotid arteries,so that the change curve of IMT can be obtained.