Authors: Deepak Gambhir; Navin Rajpal
Addresses: University School of Information and Communication Technology, GGS Inderprastha University, Dwarka, New Delhi, India ' University School of Information and Communication Technology, GGS Inderprastha University, Dwarka, New Delhi, India
Abstract: To achieve high compression ratio and good quality of compressed image, a new image compression scheme using fuzzy edge classifier and fuzzy F-transform is proposed. In the proposed scheme, fuzzy edge classifier decides the smooth or edge block, based on membership value of each block which is obtained from Gaussian function. Each smooth fuzzy block is encoded with block mean and edge block is processed using fuzzy F-transform. This encoding scheme is further decoded by applying inverse fuzzy F-transform to edge blocks and mean value to smooth block, to reconstruct the image. The output image of the decoding process shows some artefacts due to mean value of smooth blocks which is further improved by proposed Gaussian block image enhancement scheme. The experimental results show that the proposed scheme to compress the images not only improves the artefacts appearing in reconstructed image but also improves the compression ratio. The PSNR calculated in the dual fuzzy proposed method is superior than PSNR calculated in JPEG, fuzzy F-transform and fuzzy F-transform with single mean value of smooth blocks.
Keywords: fuzzy transform; FTR; Gaussian enhancement; artefact reduction; edge detection; image coding; fuzzy edge classifiers; F-transform; compression ratio; image quality; image compression; image reconstruction; image processing; PSNR; peak SNR; signal to noise ratio.
International Journal of Fuzzy Computation and Modelling, 2015 Vol.1 No.3, pp.235 - 251
Received: 01 Dec 2013
Accepted: 02 Jul 2014
Published online: 16 Jun 2015 *