The VideoLAN Forums

Discrete wavelet transform (DWT) has been heavily studied and developed in several computer graphics applications. However, when dealing with enormous data volume in computer graphics, its performance may drastically reduce.

To solve this problem, we developed a DWT shader that runs on existing GPU hardwares. We implement the convolutionary based DWT and propose an indirect addressing technique that supports various boundary extension schemes. The proposed technique also unifies the computation of both forward and inverse DWT, hence making the shaders for FDWT and IDWT are basically the same. This may simplify the future hardware implementation. The software has already been successfully integrated into the well-known JPEG2000 codec, JasPer.

To popularize the usage of DWT on GPU, we release both the full source code and binary. The software is released in two forms:
1) JasPer unofficial GPU extension (both source code and compiled binary),
2) Standalone DWT-GPU C++ Class (source code with example program).

The JasPer Software is an open-source initiative to provide a free software-based reference implementation of the codec specified in the JPEG2000 Part 1 standard. More information about JasPer can be found here. We have successfully integrated our GPU-DWT class into the JasPer software, which significantly accelerates the encoding/decoding process and is useful for handling high-resolution images.

The DWT-GPU C++ class is a free open souce software module which you can incorporate into your application and perform fast discrete wavelet transform on large data set using the parallel power of current generation GPUs.