Research On The Segmentation And Motion Estimation For The Intima-Media Of The Human CCA Wall In Ultrasound Images

The related studies have shown that the related parameters of dynamic features such as the change of the intima-media thickness(IMT)of the CCA and the two-dimensional motion mode of the intima-media complex of the CCA wall in the longitudinal and radial directions during the cardiac cycle are helpful for the early diagnosis of CCA atherosclerosis.Ultrasound imaging has the advantages of non-invasiveness,economy and safety,and is widely used in the clinical examination of the human CCA.At present,researchers generally use the technologies of image segmentation and motion estimation based on ultrasound images to perform the automatic measurement of the IMT of the CCA and the two-dimensional motion estimation of the intima-media of CCA,respectively.The robustness of the existing CCA intima-media complex segmentation algorithms needs to be further improved.However,the contrast of the ultrasound images is low,and there exist the blurred edges and the speckle noise in the ultrasound images.Due to these issues,it is very difficult to guarantee the accuracy and robustness of the automatic measurements of the IMT of the CCA based on image segmentation.Furthermore,the existing CCA wall motion estimation algorithms can be divided into two categories: the pixel level based methods and the sub-pixel level based methods.The pixel level based approaches have the disadvantage of insufficient accuracy of the motion estimation.To improve the accuracy of the motion estimation,the sub-pixel level based methods are used for the two-dimensional motion estimation of the intima-media of CCA.However,to meet the requirements of the clinical application,the accuracy and the robustness of the existing sub-pixel based CCA wall motion estimation methods need to be further improved.To address the above limitations,this thesis firstly proposes a novel denoising algorithm and a novel enhancement algorithm to suppress the speckle noise,improve the contrast and enhance the edge detail information in CCA ultrasound images.Moreover,to improve the performance of the CCA IMT measurements and the two-dimensional motion estimation of the intima-media complex of the CCA wall,this thesis also conducts in-depth research on the technologies of the segmentation and motion estimation for the intima-media of the human CCA wall in ultrasound images,and the novel segmentation and motion estimation algorithms are proposed.The main studies and innovations are as follows:1.The total variation(TV)model has attracted extensive attention due to its good performance in preserving the edge in image denoising.The related research shows that the fractional-order variational model has good performance in suppressing speckle noise in ultrasound images.Therefore,to preserve the edges while suppressing the speckle noise,this thesis proposes a despeckling algorithm for the CCA ultrasound images based on TV and fractional-order TV.Moreover,the proposed despeckling model is solved by the alternating direction method of multiplier.The experimental results show that the proposed despeckling algorithm has better performance in reducing the speckle noise while preserving the edges as compared with other advanced denoising algorithms.2.The contrast of the CCA ultrasound images is low,and there exist the blurred edges in CCA ultrasound images.Therefore,it is difficult to obtain the accurate IMT measurements based on image segmentation technology.To improve the accuracy of IMT measurements,the CCA ultrasound images need to be contrast-enhanced before segmentation to enhance the region of the intima-media while suppressing the speckle noise.Thus,this thesis proposes a CCA ultrasound image contrast enhancement algorithm based on fuzzy set theory and phase asymmetry.First,the input CCA ultrasound image is normalized.Next,the S-function is used to fuzzify the normalized image.Then,the fuzzified images are enhanced based on non-linear transform.After enhancement,the inverse transform function is used to map the enhanced image from the fuzzy domain back to the spatial domain.Finally,an edge enhancement method based on the phase asymmetry metric is used for enhancing the edges in the CCA ultrasound images mapped back to the spatial domain.The experimental results show that the proposed algorithm can achieve the better performance in suppressing the noise while enhancing the region of the intima-media complex compared with other advanced image enhancement algorithms.Thus,the proposed algorithm can lay the foundation for the subsequent accurate segmentation of the intima-media in CCA ultrasound images.3.To improve the accuracy and anti-noise robustness of IMT measurements,this thesis proposes a fully automated algorithm for segmentation of intima-media in CCA ultrasound images using improved Otsu’s method and adaptive wind driven optimization(AWDO).First,the denoising and enhancement algorithms proposed in this thesis are used for the preprocessing of the CCA ultrasound images.According to the characteristics of the preprocessed CCA ultrasound images,the region of interest(ROI)is automatically extracted from the area of the far wall of the CCA.Next,to solve the problem of insufficient anti-noise performance of the traditional Otsu’s method,the median image is obtained by applying the median filtering algorithm to the original image,then the median-average image is obtained by applying the average filtering algorithm to the median image,and the gray levels of both the original image and the corresponding median-average image are used to build the 2D histogram.Then,a line intercept histogram is established based on the 2D histogram,and the optimal intercept threshold values can be obtained by maximizing the between-class variance.Then,the proposed improved Otsu’s method combined with the AWDO is used for the multi-threshold segmentation of the intima-media.After the post process,the final lumen-intima interface(LII)and media-adventitia interface(MAI)can be obtained.Finally,the IMT can be measured automatically based on the final LII and MAI.It can be seen from the experimental results that the proposed algorithm has the lowest average absolute error among all algorithms used in the experiments,and it indicates that the proposed algorithm is the most robust and can provide the most accurate IMT measurement values.4.To improve the accuracy and robustness of the motion displacement estimation of the intima-media of the human CCA wall,an algorithm for estimating the motion displacement of the intima-media of the human CCA wall in ultrasound images is proposed in this thesis.First,select a region in the intima-media region of the far wall of the CCA in ultrasound image as the region of interset(ROI)used for block matching,the directional asymmetric search(DAS)method is used as the search method during the process of the block matching,and the sum of absolute errors criterion combined with fuzzy set theory is used to determine the similar blocks.After obtaining the best block matching position,a combination of grid slope and parabolic interpolation(GS15PI)method is used for sub-pixel estimation,and the sub-pixel position is determined.Then,the reference block from search frame-2 is used as the reference block used for block matching,the best matching position(point)obtained in the previous step and the all points in its 3×3 neighborhood need to be determined.The blocks with the same size as the reference block centered on these points are used for block matching with the reference block.After obtaining the best block matching position,the GS15 PI method is used for sub-pixel estimation,and the sub-pixel position is determined.Finally,the average of the two sub-pixel positions obtained by two block matching operations is calculated as the final sub-pixel estimation.The experimental results show that the proposed algorithm has higher accuracy and stronger robustness as compared with the existing advanced algorithm.In summary,this thesis proposes a denoising algorithm for the CCA ultrasound image using TV and fractional-order TV.This algorithm can reduce speckle noise while preserving the edge information.Next,a CCA ultrasound image contrast enhancement algorithm based on fuzzy set theory and phase asymmetry is proposed.This enhancement algorithm can suppress noise while enhancing the region of intima-media.The denoising and enhancement algorithms proposed in this thesis can lay the foundation for the subsequent segmentation of intima-media complex in CCA ultrasound images.The proposed algorithm used for segmentation of intima-media complex in CCA ultrasound images based on improved Otsu’s method and AWDO method has strong robustness and can provide accurate IMT measurement values.In addition,the proposed algorithm for estimating the motion displacement of the intima-media of the human CCA wall in ultrasound images has advantages of high accuracy,strong robustness and high efficiency that can meet the requirements of clinical applications.At the same time,to accurately estimate the change of the IMT of the human CCA wall,the automatic IMT measurement technology which consists of the proposed denoising,enhancement and segmentation algorithms in this thesis is used to automatically measure the IMT in each frame of the CCA ultrasound sequence.The quantitative evaluation results prove the effectiveness of the method for estimating the change of the IMT.