Fractal image compression offers high compression ratios and quality image reconstruction. It uses various techniques as the fractal with DCT, wavelet, neural network, genetic algorithms, quantum acceleration etc. Additionally, because fractals are infinitely magnifiable, fractal compression is resolution independent and so a single compressed image can be used efficiently for display in any image resolution including resolution higher than the resolution of the original image. Breaking an image into pieces and identifying selfsimilar ones is the main principle of the approach. In this paper, the different issues in fractal image compression as partitioning, larger encoding time, compression ratio, quality of the reconstructed image, decoding time, SSIM(Structured Similarity Index) are discussed and highlighted. The various areas for improvement as larger encoding time and PSNR are also suggested. The various parameters for evaluating the performance of these techniques as PSNR, compression ratio, encoding time, and decoding time are also suggested. Comparison of Fractal techniques for color image, texture and satellite image is done using different parameters as compression time, compression ratio and PSNR. The hybrid method which combines Fractal quad tree with wavelet and Huffman coding is implemented and compared different parameters as compression ratio and the compression time of the proposed method with the existing methods.
Many methods are available for compression of an image file. Images are usually in the form of matrices and an uncompressed image uses a huge number of bytes for storage. Its applications in various fields are quality control, remote sensing, imaging science etc. The image compression methods which are popular on the transform based coding methods like Discrete Cosine Transform (DCT), Discrete Wavelet Transform (DWT) and fractals. However, these methods have drawbacks like low compression ratio and high encoding time. The proposed hybrid technique combines DCT and fractal quad tree decomposition with Huffman encoding of fixed threshold value for color images. The results for the proposed method are displayed and compared for performance parameters as compression ratio, encoding time, decoding time and PSNR.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.