N x N DCT

• Close to KLT performance

• Appealing for HW implementation (e.g, N=8)

  • fast algorithms

Previous slide

Next slide

Back to first slide

View graphic version