What useful information i can take from the coefficient intensity maping after performing DCT transform?

enter image description here


The DCT is used in many places for its "energy compaction" properties. This is mostly useful in compression. For example, if I have a signal and I take its wavelet transform, then I can "compress" the signal by throwing away certain coefficients. These coefficients are not known before-hand. I have to look at the wavelet coefficients and then throw away all coefficients below a certain threshold to achieve some amount of compression with an acceptable loss in reconstruction accuracy.

For the DCT however, most of the energy of a signal is collected in its "low" frequency coefficients. What that means is that if you only keep the first lets say N coefficients (you don't need to store the locations of which coefficients you kept), you can get decent compression. In 2D, a "zonal" or "zig-zag" pattern is used to choose the coefficients to keep.

On a more fundamental level, the DCT is basically a linear combination of your signal as a bunch of discrete cosine waves. If you think of each cosine wave as an elementary building block, then the DCT tells you how much contribution to your image comes from a discrete cosine of a particular frequency.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.