A Study on Fractal Image Compression and Decompression by Neural Network Approaches

• Main Authors: Saik- Jen Lee, 李錫仁
• Other Authors: Koun-Tem Sun
• Format: Others
• Language: zh-TW
• Published: 1998
• Online Access: http://ndltd.ncl.edu.tw/handle/33887918163280302063

碩士 === 國立台南師範學院 === 資訊教育研究所 === 86 === This thesis mainly studies neural network technologies to fractal image compression and decompression. The fractal image compression is an important method for solving image storage and transmittion problems because of its high compression-ratio and low lost-ratio characters. However, it requires numerous comparisons and time consumption by applying traditional sequential methods. Such that, it is hard to pracitcal applications. Since many independent blocks are operated simultaneously during the fractal image compression and decompression, These operations can be processed in parallel. The neural network is a good parallel-processing mechanism, which can perform the operations by great neurons simultaneously. Then, this research is to use the parallel-processing and self-learning capabilities of neural networks to implement the numerous comparisons of fractal image compression and decompression. The simulation results show that the quality of the generated pictures and the high compression-ratio by neural networks are similar to the traditional methods, which verify the high value of our research－the neural network technologies are useful and efficient for fractal image compression.