Object Recognition Based On Fractal Image Coding

Fractal theory of iterated contractive transformation has been used in several areas of image processing and computer vision, such as compress, clouds and texture simulation. In this method, similarity between different parts of an image is used for representing of an image by a set of contractive transforms on the space of images, for which the fixed point is close to original image. Due to the self-similarity in the transformations fractal image coding can be naturally adapted for image recognition. Fractal image coding performing in object recognition is a new method, has succeed in the shape recognition and object recognition for use on binary images. In this paper, we propose it for face recognition. This method is robust to scale change, frame size change and rotations as well as to some noise, facial expression and blur distortion in the image. In this method, an unknown image is modified using the fractal code of know object. The difference between the unknown image before and after the modification is then determined using the Euclidean distance, this classification is called Fractal Neighbor Distance (FND). Fractal image coding methods infects image restoration and recognition rate, we use fractal image coding methods based on Partitioned Iterated Function System (PIFS), whose aim is image recognition, not image compression. The experimental simulation based on ORL face database shows the method proposed by this paper is effective in face recognition. So we think the image compression process equal the image feature extracting which is the point of recognition. Object recognition methods in compressed domain are reasonable. A face recognition system in DCT domain is established. The proposed algorithm can achieve higher recognition rate than the eigenface.