Medical image processing: the characterization of display changes using histogram entropy

摘要

A number of numerical processing techniques are utilized in medical imaging including spatial filtering, gradient operations, unsharp masking and histogram modification. However, it is difficult to characterize or quantify the relative nature of the changes these produce to yield an acceptable enhanced image. Histogram entropy is suggested as a useful parameter in this context. Processing techniques are described as histogram entropy-altering transformations and experimental examples are presented for computerized tomography images of the inner ear where enhancement techniques are particularly useful.