An unsharp sieve [image processing]

In some image processing and computer vision applications, it is desirable to be able to extract objects of a certain width from an image. Sifting theory, which is based on operations from mathematical morphology provides a means to extract (`sieve out´) objects up to a certain prespecified width. More often than not, the exact width of the desired objects is not known in advance. In these cases, one would like to have a sieve with `unsharp´ meshes. Objects would pass this sieve to a certain degree. In this contribution, the author presents such an unsharp sieve, which is based on local gray level operations from mathematical morphology. It takes two dimensional binary images as input and maps each pixel to the interval [0,1] proportional to the width of the object (or object part) it belongs to