Use of an adaptive filter for shape detection in microwave imaging with a focus on breast cancer detection

Detecting a target of interest is the primary focus of microwave imaging but once detected an additional useful property to have is some measure of the target shape and this study is aimed at addressing this issue. A comparison between a wideband Gaussian pulse with integration along the pulse length to a correlation approach and an adaptive filter for the purposes of seeing how finely a target shape is preserved was carried out. It shows clearly that improvements in shape preservation can be made but at the cost of higher strengths of detection. Using the filter the geometric shapes image excellently and are exceptionally clear while the cluster and enclosed shapes are recognisable and greatly improved over both correlation and pulse integration. The reflection coefficient polarity for the interface between the travelling medium and target can also be easily seen.