A Hybrid Image Codec Using Singular Value Decomposition and Adaptively Scanned Wavelet Difference Reduction Method
In this paper, a hybrid image codec for the gray scale images is proposed. It practices singular value decomposition (SVD) and adaptively scanned wavelet difference reduction (ASWDR) technique for compression. The proposed hybrid method retains the advantage of high peak signal to noise ratio (PSNR) value at a level of compression. The image is first compressed by SVD, which is a well-established approach to removing interpixel redundancies for reducing some principal components (PC) which are not significant to improve compression, and then the reconstructed image (using 100 PC) is given as an input to the ASWDR method, which is wavelet based image compression approach. In this hybrid method, ASWDR gives high compression ratio (CR), a region of interest based capability, good perceptual quality and SVD gives high PSNR value. This method is validated by some standard test images and the obtained results are compared with SPIHT and a hybrid technique using both SVD and WDR that exist in the literature. The superiority of the proposed method is well justified quantitatively and visually on basis of parameters like CR, PSNR, correlation coefficient (CC) and structural similarity index (SSIM) using the techniques. The approach obtained a PSNR value equal to 33.07 for Lena image at a compression ratio of 64:1. The percentage improvement exhibited by the proposed approach is 3.732, 10.98, and 7.83 at this compression level which is significantly higher than the comparison counterparts. Moreover, the perceptual quality metric SSIM shows an average hike of 1.92% in contrast to the techniques.
Keywords: Hybrid image codec, adaptively scanned wavelet difference reduction, singular value decomposition, image compression, DWT, lossy compression
Cite this Article
Jaiverdhan, Brijendra Budania, Barjinder Singh Saini. A Hybrid Image Codec Using Singular Value Decomposition and Adaptively Scanned Wavelet Difference Reduction Method. Journal of Open Source Developments. 2017; 4(1): 4–15p.
- There are currently no refbacks.