Image Compression Using Discrete Cosine Transform

Sura Fahmy Yousif

Department of Chemical Engineering

As our need and usage of computers are continuing to grow, so does our need for effective methods to store great amount of information. As an example, somebody who has on line catalog or uses a web page that requires a hundred of images, will need to store or transmit those images efficiently in some form of image compression. Today, several methods of image compression are available in order to reduce their storage sizes as well as using a smaller space. JPEG operation is a broadly used form of lossy image compression that related with Discrete Cosine Transform (DCT). JPEG compression reduces file size with minimum image degradation by eliminating the least important information. DCT is similar to Discrete Fourier Transform; it transforms any signal or image from spatial domain to frequency domain. This transform is important in numerous applications in science and engineering. DCT has many advantages: [1] It has the ability to pack energy in the lower frequencies for image data. [2] It has the ability to reduce the blocking artefact effect and this effect results from the boundaries between sub-images become visible. The DCT helps to separate the image into parts (or spectral sub-bands) of differing importance (with respect to the image's visual quality) during a step called quantization. In this step, the less significant frequencies are discarded, here comes the term “lossy”. After that, only the most significant frequencies are remain which are used in the decompression operation to reconstruct the original image. The restored image, as a result, will contain some distortion. But during the compression step, these levels of distortion can be adjusted. JPEG technique is used for both grayscale and color images.

Simply, the main steps of JPEG process are: [1] First the image is broken into 8*8 blocks of pixels. [2] DCT process is applied to each block from top to bottom and from left to right through a series of mathematical equations. [3] Compression process on each block is carried out through quantization. [4] The resulting array of compressed blocks is stored then in extremely reduced amount of space. [5] The original image is recovered, when desired through decompression process. In this process, Inverse Discrete Cosine Transform (IDCT) is used through also a series of mathematical equations.