Cryptography Reference
In-Depth Information
While in the flat area such as the shoulder or the face, the errors are very
small. Therefore the DCT based Intra coding has successfully compressed the
image by channelling the distortion to the areas that the human eye is less
sensitive and di
cult to detect.
(a) Original image
(b) Reconstructedimage
(c) Error image
Fig. 2.9.
Performance of DCT based image compression.
2.3 Motion Compensation Based Inter Frame Video
Coding
Although the basic principles are the same, video can be compressed at a
much higher ratio compared to still images, because in a sequence of video
frames, little changes can be found between frames. Little change means a
lot of redundant information which can be compressed effectively. Therefore
video coding typically involve two stages, the Intra coding, which refers to the
compression of a still image, as is described in the previous section, and the
Inter coding, which reduces the temporal redundancy in between consecutive
images in the video. The tools used to remove temporal redundancy is very
similar to the predictive coding as is discussed in Section 2.2.1, with the
difference that the basic unit for prediction is a macroblock (MB) which is of
the size 1616, and the prediction is based on the neighboring frames rather
than inside the same frame.
2.3.1 Frame Difference and Motion Compensated Frame
Difference
As is mentioned previously, video coding relies on removing redundancy from
the source signal. We have introduced a number of schemes to remove the
redundancy from individual images. In this section, we will show that higher
compression ratios are only achievable by removing the redundancy that exists
