Image compression with auto-encoder algorithm using deep neural network (DNN)

Abstract

An image compression is necessary for the image processing applications such as data storing, image classification, image recognition etc. Then, several research articles have been proposed to reserve for these topics. However, the image compression with auto encoder has been found for a small number of the improvements. Therefore, this paper presents a detailed study to demonstrate the image compression algorithm using the deep neural network (DNN). The proposed algorithm consists of 1) compressing image with auto encoder, and 2) decoding image. The proposed compressing image with auto encoder algorithm uses non-recurrent three-layer neural networks (NRTNNs) which use an extended Kalman filter (EKF) to update the weights of the networks. To evaluate the proposed algorithm performances, the Matlab program is used for implementations of the overall testing algorithm. From our simulation results, it shows that the proposed image compression algorithm is able to reduce the image dimensionality and is able to recall the compressed image with low loss.