Character Analysis and Recognition of the Imaging in Electrical Impedance Scanning for Breast

A new post-processing algorithm was developed for EIS image in order to improve the accuracy of diagnosis. This algorithm automatically analyzes data and recognizes the light region in the EIS image basing on the data process criterions. The threshold criterion and the modality criterion were derived from the ideal saline-tank phantom experiment. And the filter criterion came from the analysis of 240 abnormal images. Filter criterion is first used to filter some influencing data in the algorithm process. Then the threshold criterion and the modality criterion are used to recognize the light region. In addition, the algorithm was verified on a test group including 540 EIS non-processed images. The performance obtained for this test group was an 8% increase of the accuracy. The algorithm provides a clear and objective interpretation for EIS recorded images.