RESUMO
The goal of this study was to develop a calibrated on-line technique to extract as much diagnostically-relevant information as possible from conventional video-format echograms. The final aim is to improve the diagnostic potentials of medical ultrasound. Video-output images were acquired by a frame grabber board incorporated in a multiprocessor workstation. Calibration images were obtained from a stable tissue-mimicking phantom with known acoustic characteristics. Using these images as reference, depth dependence of the gray level could fairly be corrected for the transducer performance characteristics, for the observer-dependent equipment settings and for attenuation in the examined tissues. Second-order statistical parameters still displayed some nonconsistent depth dependencies. The results obtained with two echoscanners for the same phantom were different; hence, an a posteriori normalization of clinical data with the phantom data is indicated. Prior to processing of clinical echograms,. the anatomical reflections and echoless voids were removed automatically. The final step in the preprocessing concerned the compensation of the overall attenuation in the tissue. A 'sliding window' processing was then applied to a region of interest (ROI) in the 'back-scan converted' images. A number of first and second order statistical texture parameters and acoustical parameters were estimated in each window and assigned to the central pixel. This procedure results in a set of new 'parametric' images of the ROI, which can be inserted in the original echogram (gray value, color) or presented as a color overlay. A clinical example is presented for illustrating the potentials of the developed technique. Depending on the choice of the parameters, four full resolution calibrated parametric images can be calculated and simultaneously displayed within 5 to 20 seconds. In conclusion, an on-line technique has been developed to estimate acoustic and texture parameters with a reduced equipment dependence and to display acoustical and textural information that is present in conventional echograms.
Assuntos
Ecoencefalografia/métodos , Processamento de Imagem Assistida por Computador , Acústica , Calibragem , Humanos , Recém-Nascido , Imagens de Fantasmas , UltrassomRESUMO
This study addresses the value of operator-independent computer processing of ultrasonograms of the developing brain. With this aim, routine cranial ultrasonograms obtained from 39 term and preterm infants without clinical or sonographic evidence of brain damage were analyzed by five observers. The procedure, respectively, included: 1. the definition of four regions of interest (ROI), one white matter and one grey matter area on each side of the brain; 2. digitization of the sonogram data within these ROIs; 3. correction for the equipment settings, using data from a tissue-mimicking phantom as a reference; and 4. calculation of four sonogram characteristics (i.e., mean echo level, MEAN, signal-to-noise ratio, SNR, and axial and lateral correlation, CORAX and CORLAT, of the echo level co-occurrence matrix). Significant differences between both sides of the brain or a significant influence of ROI size were not found. The interobserver spread was considerable, but less than the intersubject spread. Two sonogram characteristics seemed strongly correlated in white and grey matter (CORAX and CORLAT) and another only in white matter (SNR with CORAX and CORLAT). MEAN seemed not to be correlated with any other characteristic. Furthermore, it was found that maturation equally decreases white and grey matter MEAN and, thus, hardly affects the ratio between the two. An effect on the other sonogram characteristics was only found in the white matter (i.e., an increase of SNR and a decrease of CORAX and CORLAT). Except for MEAN, the grey matter sonogram characteristics seem hardly affected by maturation. In view of these findings, we conclude that quantitative ultrasonography reveals white and grey matter maturation and, furthermore, provides a conceptional-age-independent reference (MEAN white:grey matter ratio) that might be found to facilitate the detection of pathologic brain alterations.
Assuntos
Encéfalo/crescimento & desenvolvimento , Ventrículos Cerebrais/diagnóstico por imagem , Fatores Etários , Pré-Escolar , Humanos , Lactente , Modelos Lineares , Variações Dependentes do Observador , Imagens de Fantasmas , Valores de Referência , Estatísticas não Paramétricas , Ultrassonografia/instrumentação , Ultrassonografia/métodos , Ultrassonografia/estatística & dados numéricosRESUMO
The goal of this study was to investigate the potentials of cooccurrence matrix analysis for the characterization of echographic image texture. Echographic data were obtained by one-dimensional simulations. Various data sets were generated with different number densities of the randomly distributed scatterers and with different levels of structural scattering strength. Cooccurrence matrix parameters estimated for analysis were: angular second moment, contrast, correlation, entropy and kappa. The cooccurrence matrix analysis was tuned by varying its parameters: spatial displacement of the pixel pairs, number of gray levels and size of the window. The parameters reached a saturation level at a number density of 3-5 scatterers per resolution cell (-6 dB width). Similarly, when increasing the displacement, a limit value was reached at 4-20 samples, depending on the size of the resolution cell. Using the Mahalanobis distance as a measure of differentiating between two textures, a systematic inverse relation was observed between the size of the window and the number of gray levels used in the estimation of the cooccurrence matrix. The optimal parameters to differentiate textures without structure appeared to be entropy and angular second moment. A window size of 90 speckle cells and 64 gray levels is needed for this purpose. The effective speckle size was estimated from the mean number of maxima of the demodulated echo signals. Resolved structure results in a periodicity of parameter values with displacement. The periodicity can be calculated by changing the displacement d. Optimal parameters for detecting periodicity are contrast and correlation. Analysis of the correlation between parameters showed that entropy vs. angular second moment and contrast vs. correlation are highly correlated.