Digital Signal Processing Reference
In-Depth Information
a
foreman sequence(361 frame)
46
44
42
40
38
36
34
32
H.263+ I-B-I-B
Proposed+BMS
Proposed+no BMS
Block SPIHT
SPIHT
[12]
30
28
26
0
100
200
300
400
500
600
700
800
900
rate(kbps)
b
carphon
e
sequence (101 frame)
46
44
42
40
38
36
34
32
H.263+ I-B-I-B
Proposed+BMS
Proposed+no BMS
Block SPIHT
SPIHT
[12]
30
28
26
0
100
200
300
400
500
600
700
rate(kbps)
Fig. 5.19
The PSNR of scalable DVC framework. (
a
) Foreman.qcif, (
b
) Carphone.qcif
more than 2 dB improvement than SPIHT because the dependence is used at
decoder. Also, the BMS at decoder gives more than 1 dB improvement for block
SW-SPIHT, and the scalability is still preserved. But, for Carphone sequence, the
motion searching is not so valid due to its motion being less.
