Image Compression-Based Discrete Wavelet Transform Combined with Zooming-out Algorithm

Image compression is one of the most important processes in image processing. It has numerous applications and plays a significant role in increasing transmission efficiency by reducing image size to just a few bytes, enabling the transmission of images without compromising their quality. The present study's proposed method focuses on reducing the size of an image through a compression technique. This technique involves extracting the essential information from the original image, which is represented by the low-frequency coefficients obtained through the discrete wavelet transform (DWT). By discarding the high-frequency coefficients, it can effectively preserve only the important details while significantly reducing the image size. In our approach, we employ the bi-cubic zoom-out method to resize the low-frequency coefficients, resulting in a high compression ratio. Finally, these high-frequency data are reduced by the Minimize-Final-Data method to produce compressed data (a stream of bits). Importantly, this zoomed-out image does not affect the overall image quality. We compared this proposed algorithm with both JPEG and JPEG2000 based on image quality and compression ratio. When applied to the Lena image, the proposed method achieved a PSNR of 40 dB and a compression size of 13.8 KB. However, when compared to the JPEG2000 method, the PSNR increased to 41.1 dB while maintaining a smaller compression size of 13.3 KB.