Automated image interpretation of digital chromatic images using multivariate statistics

In recent years, as the amount of data being gathered for biomedical imaging applications has begun to overwhelm human processing, an emphasis has been placed on automated image interpretation. The principle drawback of classical segmentation and shape analysis approaches to automated image interpretation is that they lack sufficient sensitivity to address skewed or overlapping image and object points in feature space. In addition, most approaches require a high degree of human interaction in order to be effective. The focus of this research is the development of a set of computer vision algorithms capable of automated segmentation, shape analysis. and object recognition in digital chromatic images using multivariate statistics.