Untitled Page

In the domain of Object Recognition, it exist two main systems, Semi-Autonomous Object Recognition, and Autonomous Object Recognition. The first deals with the case of the application of object detection using A Priori information, describing its general model or different classes. The second is based on no A Priori information, in other words “the signal defines it self”. Of course these two points of view for object recognition are widely connected. For an example, an “object feature” that is extracted from an object prior to identification according to an A priori model, is the same feature that can be used to compare two objects in a totally unsupervised environment. In our approach for Object Recognition, we followed an object-oriented manner in which the Object is a product of a multidimensional process and is the same everywhere. Object-oriented concept, as a programming technology is very used in our modern days. In the Object-Oriented theory any Object of any form can be considered a descendent or a parent to other objects. We did try to invest this abstract point of view of objects, by the principle of the relativity of resemblance. That is because this relativity is what is permitting the construction of any abstract hierarchy. For that, we build a mathematical measure that is compatible with the relativity principle, in the sense that any object resembles to another in multiple correspondences.