My Calendar

Tuesday 19 February 2008

PhD Works: Algorithm, Architecture and Hexagonal Plot

Motion estimation(ME) is a part of components required in designing an efficient encoder. It's always the most complicated part of the system, and can absorb huge system resources and it also the computational most demanding algorithm for a video encoder (60 - 80%) of total computational time. Just imagine! how important the ME for video codec.

Basically, in this PhD works it can be divided into the following parts: Motion Estimation (ME), Motion Compensation (MC) (including search algorithm, cost calculation, mode decision, store/retrieve data, and signal and controlling and etc. - check your POS)
The following figure, respectively is the spectrum of issue for the ME algorithm and ME architecture in order to find the room of contribution in this area.





This is very interesting finding in terms of how to make a better comparison for any design. It is used to visualize the characteristics of ME hardware architectures. I need to explore it more, for better publication in future. Thanks for the paper author for your great ideas (Chen et al. 2006).

~May Allah help me. PhD is a journey of how to be a good Muslim~

No comments: