Cryptography Reference
In-Depth Information
2.3.3 Motion Compensation
During the motion estimation, matched MB in the reference frame is used as
a prediction of the current MB, and their difference forms a low-information-
content prediction-error block. This prediction error will be compressed the
same way as the DCT based Intra coding. Its corresponding motion vector
is further compressed using DPCM coding. Both the motion vector and the
prediction error are sent to the decoder, see Fig. 2.13(a). In the decoder, the
motion vector is then used to decide the displacement in the reference frame,
and the displaced MB in the reference frame will be extracted and used as
the prediction for the current MB. In the meantime, the prediction error is
Intra decoded, and added to the predicted MB to reconstruct the final MB.
Fig. 2.13(b) shows the process of motion compensated decoding.
Input video sequence
Input video sequence
+
+
Intra
Coder
Intra
Coder
+
+
Prediction
error
Prediction
error
_
_
Intra
Decoder
Intra
Decoder
Motion
Estimation
Motion
Estimation
Motion vector
Motion vector
MV
Coder
MV
Coder
Predictive
coded output
Predictive
coded output
Predicted picture
Predicted picture
Motion
Compensation
Motion
Compensation
+
+
Reference
Frames
Reference
Frames
(a) Motion compensated predicative encoding
Compensated
difference
Compensated
difference
Intra
Decoder
Intra
Decoder
Predictive
coded input
+
+
Decoded
frame
Decoded
frame
DPCM
Decoder
DPCM
Decoder
Reference
Frame
Reference
Frame
Motion
vector
Motion
vector
(b) Motion compensated predicative decoding
Fig. 2.13.
Motion compensated predicative encoding and decoding.
