HYBRID IMAGE COMPRESSION TECHNIQUES USING DWT AND NEURAL NETWORKS

One of the most important methods for cutting the costs of digital image transmission and storage is image compression. In order to obtain large compression ratios and good image quality, this research proposes a hybrid image compression technique that combines the benefits of Neural Networks (NN) and Discrete Wavelet Transform (DWT). The input image first has to be divided down into segments at various frequencies using DWT. After being quantized, the sub-bands are put into a neural network to be further compressed. The neural network is trained to produce compressed representations with minimal data loss and to understand the statistical characteristics of the image's sub-bands. Next, a lossless or lossy compression algorithm is used to encode the compressed image data, which is then either saved or transferred. According to experimental findings, the suggested hybrid compression method performs better in terms of compression ratio and image quality than conventional DWT- and NN-based compression methods. Furthermore, by varying the neural network design and the compression settings, our method is adaptable to various compression requirements.