A LOW-COMPLEXITY DCT APPROXIMATION FOR IMAGE COMPRESSION WITH 14 ADDITIONS ONLY

摘要

The discrete cosine transform (DCT) is one of the basic tools of real time image and video compression systems. In this paper, we proposed a low-complexity 8-point discrete cosine transform. The proposed 8x8 transformation matrix requires only adders, thus avoiding the need for multiplication and bit shift operations. The new class transform requires only 14 additions, which highly reduces the computation complexity with negligible degradation in peak signal to noise ratio (PSNR). The proposed transform outperform the existing methods by a 0.81dB improvement in the average PSNR and a 1.19dB improvement in the peak PSNR.