Performance Comparison of TS and Haar Wavelet Transforms for Real-Time FPGA Implementation of Image Compression

摘要

As demands for speed and quality in the transmission of digital image data continue to grow, improved methods for image compression become necessary. Wavelet analysis is a mathematical tool useful in digital computer applications such as image compression and real-time video transmission. Integer Wavelet Transforms present certain inherent benefits such as computational speed, simplicity, compressibility, and reconstructed image quality. Simple Wavelets such as the Haar Wavelet are relatively easy to implement in digital hardware. Due to its extreme simplicity, however, the Haar Transform renders reconstructed images that contain small but noticeable errors. A more robust wavelet, the TS Wavelet, provides better reconstructed images while requiring only slightly more complex computations. This paper briefly discusses wavelet transforms and how they are employed in image compression techniques. Specifically, the focus of this work is on the performance of the TS and Haar Transforms' image quality and integrity. The comparative performance study is carried out using FPGAs for the hardware implementations of the image compression schemes.