الفهرس 
Color spaces Motion AnalysisOnly 14 pages are availabe for public view
 |  | from | 202 |  |  |
| | | from | 202 | | |
Abstract
It is more important to build video compression systems that give higher compression
ratios with good isual quality in the same time. Beside the higher processing speed is an
important factor in those compression systems. Those compression systems are used in
high definition (HD) video devices. broadcasting more television channels on the satellites
and mobile phones which can deal with the multimedia.
The ideo compression operation depends on removing the temporal redundancies due to
motions between the successive video frames. Then spatial redundancies within the frames
an: rernox ed. Finally the entropy operation is appl ied to get the compression operat ion.
The temporal redundancies are removed by applying Motion Estimation (ME) and
Motion Compensations (Me) operations to get motion vectors and residual frames. The
residual frames have lower energies than original frames so it is easier to be compressed.
The spatial redundancies are removed by the transformations operations as di crete
wavelets transform OWl. Finally the statistical redundancies are removed at the entropy
stage. In the entropy stage the modem method called Embedded Block Coding with
Optimized Truncation (EBCOT). used in the image compression standard JPEG 2000, is
applied. The EBCOT method gives higher compression ratio with good visual quality.
Designed MATLAB tool programs are used in the motion analysis and the
transformation stages. Those programs are applied on standard video frames length 50
frames and of size 128xJ28 pixels. To ensure the accuracy of those programs the inverse of
those programs are applied in the reverse direction. The original frames are compared with
resulted frames and lose was small.
The EBCOT MATLAB tool programs are applied on the OWT transformed residual
frames form the motion and transformation stages. Then the compression ratio and the
visual quality are calculated. The compression ratio is witb range 40% 10 75% with
decreasing the visual quality between 34% to 60 % respectively.
The main problem of the EBCOT algorithm is that it consists of t\VO complicated stages.
The first stage called Context Formation in which the contexts decisions pairs are
generated. The CF deals with bits of during frame processing so it needs longer processing
period. The second stage is Arithmetic Coder which uses contexts decisions pairs to
generate the compressed data. But also the AC is too complicated and need long processing
period. The used Arithmetic Coder is famously known as Context Adaptive Arithmetic
Coder (CABAC). To use the CABAC algorithm as an entropy method it is required to
design high speed architecture to overcome its long processing period
This thesis presents high speed design for the EBCOT algorithm on field programmable
gate array (FPGI\). So the proposed design can deal with the digital video signals. The
simulation is done on chip of type Yirtex xcv200e of Xillinx Company. This chip has a
feature high speed performance and high capacity.
The simulation results indicate that the design can operate at frequency 93 MHz after
place and routing and can process a video frame of size 128x 128 pixels in 7.8 m St:C. Also
from the comparing the proposed design with some modern designs, this design achieves
higher processing rate with 51 % and decreasing the deJay time with 56.6 % but with
increase in the used gate with 37% and more power consumption with about 24%.