An Improved Image Segmentation Algorithm Based on MET Method

Image segmentation is a basic component of many computer vision systems and pattern recognition. Thresholding is a simple but effective method to separate objects from the background. A commonly used method, Kittler and Illingworth's minimum error thresholding (MET), improves the image segmentation effect obviously. It’s simpler and easier to implement. However, it fails in the presence of skew and heavy-tailed class-conditional distributions or if the histogram is unimodal or close to unimodal. The Fisher information (FI) measure is an important concept in statistical estimation theory and information theory. Employing the FI measure, an improved threshold image segmentation algorithm FI-based extension of MET is developed. Comparing with the MET method, the improved method in general can achieve more robust performance when the data for either class is skew and heavy-tailed.