Light network high performance discrete cosine transform for digital images

摘要

This paper proposes a new high-performance Discrete Cosine Transform (DCT) for image and signal processing applications. Generally, DCT has been developed in a wide range of applications such as image compression and motion detection. However, the high complexity and required processing have effectively minimized the practical utilization of DCT. A New mathematical derivation for DCT is achieved in this paper with the purpose to reduce its complexity and speed up its processing. The new fast DCT version is mainly based on removing coefficients redundant calculations. The new developed model has shown a potential performance in comparison with standard DCT. Moreover, the proposed model has a high quality that has been investigated based on Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR). Fast DCT requires only 0.15%–6% processing time in comparison with traditional DCT.