This work investigates some motion estimation algorithms for video compression and it presents results from software implementations for the investigated algorithms. These results were used to define the motion estimation architecture presented in this paper. The designed architecture uses the 4:1 Pel Decimation algorithm with SAD, in blocks with 16x16 samples. The search area was defined as 64x64 samples. The synthesis results show that this architecture is able to processes more than 60 SDTV frames (720 x 480 pixels) per second. This result shows that the designed architecture is able to compress SDTV videos in real time.
|