tailieunhanh - Recent Advances in Biomedical Engineering 2011 Part 17

Tham khảo tài liệu 'recent advances in biomedical engineering 2011 part 17', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | A 3D Graph-Cut based Algorithm for Evaluating Carotid Plaque Echogenicity and Texture 629 0. 0. c 0. 0. 0 2 4 6 8 10 n Fig. 4. Regularization parameter function depending on the number of observed pixels. Feature Extraction The volume field F describes the value of the Rayleigh parameter across the VOI containing the carotid plaque and it may be used to compute local intensity and textural indicators that characterize the components and tissues of the plaque. Traditionally plaque characterization is based on statistics computed from the observed noisy pixels. Here instead of computing these indicators from the noisy data the characterization is based on statistical estimators depending on F containing the Rayleigh echogenicity intensity parameters throughout the whole plaque. The echogenicity and texture measures are derived from statistical estimators for the mean fp p median GSM fv p standard deviation f p and percentile 40 fP40 p depending on fp where p i j k . These statistics derived from the Rayleigh distribution Abramowitz et al. 1972 are given by f p -fn . fv p J2log 2 fp f x j4-2 fp fP40 x 1 - e 2 15 For each 3D reconstructed volume F four 3D matrices Fp Fv Fo and FP40 were built containing the values of fp p fv p fo p and fP40 p respectively at each location p. By averaging the elements of these 3D matrices global indicators about the plaque can be computed. Thus the global echogenicity of the plaque is given by the mean of Fp Fp and the percentage of pixels with grey-scale lower than 40 which is known to be a relevant feature is given by FP40 . Additionally the intensity variability is given by Fv and FO . Local Characterization by use of Graph-Cuts The global characterization of carotid plaques described in the previous section by averaging the statistical estimators in 15 despite its unquestionable usefulness may not be 630 Recent Advances in Biomedical Engineering enough for a correct assessment of plaque vulnerability especially in cases .