Fast Coding of Haar Wavelet Trees
Dylan Tarter, Brian S. Nutter · 2022
Compression methods that focus on speed and computational efficiency are of particular interest when acquiring real time data. The Backwards Coding of Wavelet Trees (BCWT) algorithm by Guo et al. offers high computational efficiency and high peak signal-to-noise (PSNR) ratios at a high speed. In this paper, the BCWT algorithm is reworked using a Haar Wavelet Transform, rather than Debauchee 9–7, to provide even faster and less complex code at the cost of lower PSNR. While in typical image compression applications this compromise is unfavorable, some real-time data acquisition systems require such improvements and are not significantly impacted by the reduced clarity.