Skip to content

Discrete Cosine Transform (DCT)

ashumeow edited this page Dec 7, 2014 · 1 revision

DCT is composed of 8 standards namely DCT-1, DCT-2, DCT-3, DCT-4, DCT-5, DCT-6, DCT-7 and DCT-8.

The DCT-1 mainly discusses on the orthogonal power of the eigenvectors.

The DCT-2 deals with the approximation of eigen vectors. While dealing with scaling towards DCT-1 to DCT-4, the eigenvectors becomes orthogonal. The boundary conditions give second order accuracy around the center points.

The DCT-2 is where its magic is applied in JPEG image compression by means of 8X8 block.

The DCT-3 gets defeated by DCT-4. The DCT-4 focusses on orthogonal side and matrix factorization, where it also reveals the usage of Discrete Sine Transform (DST).

The DCT-5 and DCT-6 have a great coding gain when compared to DCT-2.

The boundary conditions can be easily overlapped in DCT-7 and DCT-8 when compared to DCT-4. Thereby, It has been analyzed that Discrete Cosine Transform (DCT) is more efficient.