Design of Novel FFT Based Image Compression Algorithms and Architectures

Image Compression is a pivotal step in Digital Image Processing area. Image Fourier Transform is long established algorithm which converts an Image from spatial domain to frequency domain. New Algorithms for FFT, Quantization and their inverses have been developed using Algorithmic State Machine Charts. Subsequently, novel Architectures have been designed using massively parallel and highly pipelined circuits. These Architectures have been coded using Verilog Hardware Description language conforming to RTL coding guidelines, simulated and tested for color images. The quality of the reconstructed images is better than 35 dB. The reconstructed images are indistinguishable from the original images. The target clock rate is 50 MHz and compression achieved is greater than 15.