Analysis of motion estimation algorithms in video

journal of Artificial Intelligence in Electrical Engineering

دوره 10، شماره 37، شهریور 2021، صفحه 1-6   

نوع مقاله: Original Article

نویسندگان

Dadvar Hosseini¹؛ Mahdi Noshyar²؛ Saeed Barghandan^{* 1}؛ majid ghandchi¹

¹Department of Electrical Engineering, Ahar Branch, Islamic Azad University, Ahar, Iran

²Associate professor, Department of Electrical Engineering, University of Mohaghegh Ardabili,Ardabil, Iran

چکیده

One of the most time-consuming parts of video encoding is motion estimation. Motion estimation is an important part of video encoding that improves the amount and accuracy of compression by exploiting the temporal similarity between successive frames. This action is the main cause of maximum compression in the video, which imposes the minimum quality loss on the video. In order to be able to compensate the movement, it is necessary to first find the value of the movement direction, which is called movement estimation.
By doing motion estimation, instead of sending repeated information at consecutive times, the difference between consecutive frames is encoded and sent in different ways. Generally, a motion estimator in video encoding includes fractional pixel motion estimation (FME) and integer pixel motion estimation (IME). The FME method provides better performance with the cost of computational complexity and estimation time on the encryption process.

کلیدواژه‌ها

motion estimation؛ block matching؛ sequence؛ frame؛ pixel