Title: Fractal Image Compression Using Dynamically Pipelined GPU Clusters
Abstract: The main advantage of image compression is the rapid transmission of data. The conventional compression techniques exploit redundancy in images that can be encoded. The main idea is to remove redundancies when the image is to be stored and replace it back when the image is reconstructed. But the compression ratio of this technique is quite insignificant, and hence is not a suitable candidate for an efficient encoding technique. Other methods involve removing high frequency Fourier coefficients and retaining low frequency ones. This method uses discrete cosine transforms(DCT) and is used extensively in different flavors pertaining to the JPEG standards. Fractal compression provides resolution-independent encoding based on the contractive function concept. This concept is implemented using attractors (seed) that are encoded/copied using affine transformations of the plane. This transformation allows operations such as, skew, rotate, scale, and translate an input image which is in turn is extremely difficult or impossible to perform in JPEG images without having the problem of pixelization. Further, while decoding the fractal image, there exist no natural size, and thus the decoded image can be scaled to any output size without losing on the detail. A few years back fractal image was a purely a mathematical concept but with availability of cheap computing power like graphical processor units (GPUs) from Nvidia Corporation its realization is now possible graphically. The fractal compression is implemented using MatLab programming interface that runs on GPU clusters. The GPUs consist of many cores that together give a very high computing speed of over 24 GFLOPS. The advantage of fractal compression can have varied usage in satellite surveillance and reconnaissance, medical imaging, meteorology, oceanography, flight simulators, extra-terrestrial planets terrain mapping, aircraft body frame design and testing, film, gaming and animation media, and besides many other allied areas.
Publication Year: 2014
Publication Date: 2014-01-01
Language: en
Type: book-chapter
Indexed In: ['crossref']
Access and Citation
Cited By Count: 4
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot