Haralick texture features extract frequencies of local spatial variations in signal intensity, quantifying pixel relationships within regions of interest. In this study, Haralick texture features were used to differentiate biopsy-proven, cancerous lesions from non-suspicious regions in low-field MR images. The results demonstrate consistency in texture measures for cancerous regions compared to non-suspicious, where Energy and Homogeneity are elevated while Contrast and Correlation are reduced within cancerous regions compared to non-suspicious regions.
This abstract and the presentation materials are available to members only; a login is required.