Wavelet Transforms and Applications to Still Images and
Video Compression
| hcycon | fhtw-berlin.de |
|---|
Discrete Wavelet Transforms (DWT), differently from Discrete Cosine Transform (DCT), have in-build space-frequency localization properties that enables image coding without blocking artifacts. This implies that DWT has higher rate-vs-distortion performance than DCT has for still image compression. Therefore, a DWT coder was selected for for the new JPEG 2000 standard.
For video compression at bit rates less than 64 kbps, the performance of our DWT coder is superior to the performance H263++ coder proposed as reference by ITU-T.
The talk has two demos, a high-performance still image wavelet codec and a real-time video wavelet codec that were prototyped under support from Deutsche Telekom Berkom GmbH.