HONN approach for automatic model building and 3D object recognition

摘要

This work presents a method for automatic model building from multiple images of an object to be recognized. The model contained a knowledge have been computed during learning phase from a large 2D images of an object. This knowledge is the invariant features includes to an object it self, and is extracted by using Higher-ordered Neural Network (HONN) structure. In the recognition phase, an independent viewpoint 2D stereo images of object are taken and the invariant features are extracted from it, and compare to the models stored into the data base. Both model and recognition algorithms are be tested practically to get a optimal compact model with acceptable recognition rate.