Object recognition using wavelets, L-G graphs and synthesis of regions

This paper presents a recipe of methodologies for object recognition using wavelets, local-global (L-G) graphs and a region synthesis approach. In particular, the wavelets rearrange the shape of an object for reaching a desirable size, the L-G graphs represent the shape, color, texture and location of each image region obtained by an image segmentation method and the synthesis of the regions that compose an object is achieved by synthesizing their graph representations under certain neighborhood criteria. The synthesis of the regions is based on the L-G modeling that compares a set of L-G-based object models sitting in an L-G database. The methodology is accurate for objects existed in the database and it has the ability of “learning” new L-G patterns associated with objects that do not exist in the database. Illustrative examples are also provided.