The EFRM: An Intermediate Image representation Scheme

Specifies a method for the creation, from an image, of an Enhanced Facet Region Model (EFRM), a powerful intermediate representation scheme which can form the basis for a variety of applications within the computer vision field. The desire to construct such a model has grown from observations on the failings of edge detection algorithms to provide a rich enough description of image events from which to execute higher level analysis. The ERFM is built on the basis that surface orientation provides a robust and useful metric upon which to form an intermediate image representation. >