Design and Implementation of Image Compression Encoder using Orthogonal Approximation DCT

Vanitha R, K S Gurumurthy

approximation discrete cosine transform (ADCT), compression, discrete cosine transform DCT), high efficiency video coding (HEVC, structural complexity.

Image Compression is usually carried out using discrete cosine transform (DCT) because compressed image using DCT will take less memory to store the image and quality of the image will be good compared JPEG and HEVC. But, in this work an attempt is made to achieve compression using Approximation DCT (ADCT). ADCT is useful for reducing its computational complexity without affecting its coding performance. It provides better image and video compression compared to the DCT. ADCT is orthogonal and it has lower structural complexity compared to DCT. The unique feature of the ADCT is that it could be configured for the computation of the 32 point ADCT or for parallel computation of two16 point ADCTs or four 8 points ADCTs. It has many advantages in terms of orthogonality, structural simplicity and lower computational complexity. The proposed ADCT is implemented using Verilog and Simulated by ModelSim and synthesized by Xilinx ISE 9.1i. Results are compared with 16 point ADCT with 16 point DCT implementation. The target device is XC5vtx330t-2ff1738. The 16 point ADCT implementation results in a saving of 28.37% IOBs and 63% of LUTs, compared to existing 16 point DCT implementation.