The Study And Implementation Of Image Retrieval Based On Fractal Encoding

As the "Big Data" era is approaching, the primary difficulty that we are facing is no longer the lack of information but how to take full advantage of the huge resource.Therefore, a efficient and intelligent retrieval system has become the focus of the current research.With digital equipments into sight, increasingly convenient access to image information makes its way into all aspects of life and has become the best important property of each organization. However, a current image retrieval system can’t meet the people quickly efficient retrieve information needs. Consequently, as this research background, paper according to the existing problems has the related research on extraction algorithm based on fractal coding features and fractal image coding retrieval algorithm, aims at improving the retrieval performance and demonstrates this thesis. The main work is as follows:1. Full investigation and research is conducted for the development of image retrieval system and its key technology based on content is also studied, among which feature extraction algorithm based on fractal image retrieval algorithm, the principle, the key factors influencing the performance and existing problems is specially analyzed.2. For the current feature extraction based on fractal image coding algorithm analysis, the main problem has been found that the image feature extraction fractal encoding algorithm can not fully replace the original image, and the image feature here is the fractal image coding files. Therefore, judging the fractal coding file will be able to replace the original basis is fractal coding reconstruction image distortion rate. Accordingly, combined with Fisher block classification and correlation coefficient an improved fractal coding algorithm is proposed. And this improved fractal coding algorithm will reduce fractal coding reconstruction image distortion rate. The final simulation results verify the algorithm reduces the current fractal coding algorithm reconstruction image distortion rate.3. For the current fractal-based image retrieval system, all images when storage needs through fractal coding algorithm fractal coding files, it can’t be able to achieve online retrieval and leads to a lack of flexibility. Hence, this paper presents an image retrievalalgorithm based on two-dimensional entropy and fractal image. Instead of the form of fractal coding files stored in the database, the method in this paper simply segments the image after screening the database and verifies its actual use in retrieval system. Experiments show that the algorithm performs well in online real time search.4. Based on the novel image retrieval system proposed in 3, we realize the algorithm on MFC and Cx Image library, and the system can not only achieve the function of image retrieval, but also can be used for comparison with other methods.