Object Recognition Based On Fractal Neighbor Distance And Quan-tree Partition Coding

Fractal theory has achieved a series success in the image processing, especially in fractal image coding. It is also improved effective in object recognition.Fractal image coding can approximate any given image by capturing intrinsic self-similarities within the image. Due to the self-similarity in the transformations fractal image coding can be naturally adapted for recognition. In the paper, it presents a new way based on fractal neighbor distance for object recognition and present a detailed method and procedure.Fractal neighbor distance gives a quantitative measure of the input-output characteristics of the fractal code of an encoded image. The fundamental mechanism behind this recognition scheme lies in the uniqueness of the attractor of a fractal code. The attractor has the invariance to image translation, rotation, scaling and illumination.In the coding scheme, this paper uses the coding method based on quad-tree partition, which increases the speed and veracity of the coding and is useful for object recognition. There are various flags during the quad-tree partition coding. In order to get the best coding, this paper alteres the flags, which include maximum recursion depth, minimum recursion depth, domain pool type, scaling bits, offset bits, number of iterations.The experimental simulation based on ORL face database shows that the method proposed by this paper is effective in object recognition.