Quantitative ultrasound tissue characterization using texture and cepstral features

Various researchers have used texture analysis to perfonn ultrasound tissue characterization with mixed results. Others have used spectral parameters to classify tissue. Several groups have used a feature attributed to the mean scatterer spacing (MSS) as a discriminating feature for tissue classification. We have previously shown that the locations of peaks in the complex cepstrum (PCEP) provide a good estimate of MSS in phantom and simulation studies. In this work, we show that by combining PCEP with texture-based features, we are able to distinguish between normal and hepatitis patients with a high degree of accuracy. In our preliminary analysis of chronic hepatitis patients, we have achieved reasonable classification performance (area under the ROC curve, A(z)= 0.86 +/- 0.04) using just two features (location of peak in the complex cepstrum and the entropy of the co-occurrence matrix). We are continuing our analysis to develop texture, spectral and cepstral features that provide improved, machine independent, tissue classification using multivariate analysis.