Abstract
Identification of loss of the gray-white matter interface (GWMI) is important for the diagnosis of acute stroke in unenhanced CT. However, the identification is difficult due to the intrinsic presence of quantum noise on CT images. As a result, low sensitivity of detection of loss of the GWMI becomes a major issue in unenhanced CT. To cope with this issue, we have previously developed an adaptive partial median filter (APMF). In order to validate the usefulness of the APMF, we have evaluated the performance of the proposed filter using simulated images and clinical images. This report first describes the outline of the APMF and its performance. Next, the practically clinical usefulness of the APMF is presented based on experimental results.
