Directional analysis of images in scale space

The authors propose a computational technique for the directional analysis of piecewise linear patterns in images based on the notion of zero crossings in gradient images. A given image is preprocessed by a sequence of filters that are second derivatives of 2-D Gaussian functions with different scales. This gives a set of zero-crossing maps (the scale space) from which a stability map is generated. Significant linear patterns are detected from measurements on the stability map. Information regarding orientation of the linear patterns in the image and the area covered by the patterns in specific directions is then computed. The performance of the method is illustrated through applications to a simple test image made up of straight bar patterns as well as to scanning electron microscope images of collagen fibrils in rabbit ligaments. The method has significant applications in quantitative analysis of ligament healing and in comparison of treatment methods for ligament injuries.<
>