Research On Algorithm Of Fractal Image Coding Based On Regional Classification

In various multimedia services and digital communication and other fields of application, image coding is crucial technical one. In the 1990s, fractal image coding as a new compression coding method, it began to rise the integral theory is based on image self-similar foundation. Because a lot of self-similar geometry exist in nature, therefore, in the compression process should not only consider the image similarity between the regional and global, regional and local more to consider the similarity between the object, so it can handle for natural image compression condition speaking, fractal coding technology will have a very wide range. At present, the fractal coding deficiency and: decoded image effect also insufficient ideal, encoding speed needs improving, produce better compression haven't reached the level in theory. But we through theoretical analysis know, encoding speed slow is the practical fractal coding can't the biggest obstacle, but also can be optimized fractal image coding, the potential advantage with a lot of room to rise.The study in this paper mainly as follows:(1) The detailed analysis of the fractal coding encoding and decoding algorithm, and the realization process of the basis of the coding realize, verified by the experiments sub-block division size and father block isometric transform number two big factors on the influence of fractal coding performance.(2)The detailed analysis of the human visual system related characteristics and visual characteristics and image texture, and further discusses the correlation of the visual characteristic applied to image coding the improvement scheme.(3) By using the analysis of image region classification to achieve in guarantee under the premise of subjective image quality fractal coding time performance. Focuses on the comprehensive consideration of the various visual influence factors to the method of image region, according to this article classified improved algorithm of ideas, can be a very good will image segmentation sub-block is divided into smooth area, texture area and edges area.Through the final experiment results show that this analysis of the image quality and the encoding time from improve fractal coding the effectiveness of the new plan. Through the use of improvement is to realize this reduction strategy of code, that is, from human vision remains characteristic embarks, ensure the quality of image cases, reduce the quantity, and matching codes corresponding full search fractal algorithm in decoded image comparison, subjective and objective quality under the precondition of no apparent decline, coding time greatly decrease.