Just Noticeable Difference Based JPEG Image Compression

In this study we present about JPEG lossy image compression and decompression.The fastest growing of digital technology and information exchange makes an image compression become an issue and essential.Nowadays we can use multimedia in different areas for different purposes such as in hospital,banking,insurance,aviation,science and technology,education centers,mass media,gaming industry and like.Image compression is an important topic in those areas and it plays a vital role in digital image processing.Even if bandwidth and storage hardware technology become advanced,still the amount of information encoded in an image is very huge.However,the existing bandwidth of network and storage of memory devices restrict the rapid development of multimedia products.To save more hardware space and transmission of bandwidth image compression is essential for reducing the redundancy of data.Compressing an image is used to minimize the amount of data stored in storage devices as well as the data transmitted time over the web.To satisfy the need of high quality image compression methods have been developed.One of this is JPEG compression standard and it has also compressed and decoded by different methods and different algorithm.There are different methods of algorithms in compressing JPEG image compression.Basically,it is classified into lossless and lossy image compression.The method of JPEG image compression is used according to the importance and the quality of output images.If the image compression application is expected to generate a very high-quality output without any loss in reliability,it's possible to use lossless compression method.This method is used where a high degree of accuracy is necessary and logically it doesn't loss information.In JPEG lossy image compression algorithm an image is split into eight by eight blocks and applied 2D Discrete Cosine Transform(2D DCT)which extracts spatial frequency information from the spatial amplitude samples.Just Noticeable Differences(JND),which reveals the visibility of our Human Visual System(HVS),is very important for image and video coding.Researchers on cognitive science point out that the HVS is highly adapted to extract the repeated patterns for visual content representation.JND method acts an important role to removing perceptual redundancies in image and video compression.To achieve better perceptual quality and more effective image compression,we propose Just Noticeable Differences(JND)Based JPEG models for images.This research used pixel-wise JND model with the combination of basic JPEG lossy image compression algorithm.We compute mean JND value for each eight by eight block and adjust each pixel by its corresponding mean JND value.The dataset,which contains the images used in the study is collected from different webs and select appropriate image for experimental tests.Experiments were done using"MATLAB R2015a" version.JPEG image compression and JND Based JPEG compression are experimented in order to compare the experimental results.The subjective viewing tests conducted that the perceptual quality improvement has been confirmed.Besides,for objective quality evaluation,we compare the bit rate and PSNR of grey scale images.We achieve average improvement of 0.11 dB in the objective coding quality measure(PSNR)on average and the reduction in average bitrate of 0.05 bpp over the ten images which have been tested.In this study,an experimental result shows that the proposed method is effective for JPEG image compression.