Development of Wavelet Based Image Compression Approach Using Modified SPIHT Algorithm

Abstract

In case of images & videos, the encoding philosophy is substantially different as compared to that of speech. The very basic difference between image & speech is that, the speech signals are one dimensional, whereas the images are two dimensional in nature. Modern day images can have the resolution in the range of ‘1024 x 1024’, this means there will 1 Mega Pixels & for color images, 1 pixel is represented by 24 bits (8 bits for each pixel component of red, green & blue), that means there are 24 Megabits in a single still image. In videos such still images are changing at 30 frames per second, so the basic bit rate going to be 30 times of 24 Megabits, this leads to the explosion in information. This demands compression of images at significant extent. In this work, image compression is achieved by using transformation based on modified lifting wavelet structure followed by quantization based on modified set partition in hierarchical tree (SPIHT) algorithm. This modification in SPIHT algorithm helps to reduce the bit stream length of compressed signal to significant extent as compared to conventional SPIHT algorithm. The proposed architecture is developed with Verilog language and simulated with modelsim simulator and synthesized with XST(VHDL/Verilog) tool of Xilinx software under Virtex6-XC6VLX240TD environment. The comparative analysis of proposed system with existing system shows that, there is significant improvement in compression ratio & PSNR.