Study the Effect of Noise on Compressed Images Used in Smart Application Based on JPEG Standard

Abstract

Recently a lot of smart applications based on using a data set, in the most of the cases, the data set is images, like in smart systems based on detection, recognition and auto decision, also in the systems based on data transmission and smart networks, according to those applications the most critical problem is our ability to save this data from the noise effect, which really could create wrong message or makes our data unclear for proving and analysis, however using data in its original format could take long time, which will consume our storage capacity, the bandwidth usage, processing resources and the energy used for the operation, this will lead us to use a kind of compression that gives us the best solution for all the drawbacks mentioned before. The JPEG compression gets a lot of attention in this term, since its produce a high-compression ratio with reconstructed image close to the original one, due to using DCT transform, which give us a good representation of the image in the frequency domain, however with all this benefits the JPEG standard is so sensitive to the noise effect, since the encoded data related to each other, its look like a related chain, so the smallest perturbation causes a tremendous collapse in terms of decoding (reconstruction of image), in this paper we are going to test and study the data sensitivity to the channel noise based on transmitted using JPEG compression, which allows us to offer efficient techniques in terms of restoration or data correction.