 |
First, we transform the original data using DCT signal transformation for the compressing data.
After changing, the actual data consists of transformation coefficients. We define energy for each coefficient because each coefficient component has its energy.
Next, we fond k, which determines how many coefficients represent the energy in the data in terms of energy concentration.
To achieve a higher compression ratio, we ignore the non-significant components other than the k-dominant ones.
Consequently, in our compression method, it is possible to achieve a higher compression ratio if the number of dominant coefficients is sufficiently smaller.
We need K-dominant components and their indices only as approximation datasets. In the reconstruction steps, their indexes are used to make forms the same size if the original data.
|
IEEE/ICACT20220021 Slide.07
[Big Slide]