摘要 |
The disclosure relates to a method for estimating, for a set of pixels defined in an initial frame I 0 , a function of the positions of the set of pixels in a reference frame I n . According to the disclosure, said method comprises at least one iteration of the following steps, for a current frame of the sequence:
- obtaining (41) a matrix P k,k +1 representing, for each pixel of a previous frame I k , a probability of the pixel to be at different positions in the current frame I k +1 , the previous frame I k preceding the current frame I k +1 in the sequence;
- determining (42) a vector E k,n of function of positions of the set of pixels in the previous frame I k , by multiplying said matrix P k , k +1 with a vector E k +1 ,n of function of positions of the set of pixels in the current frame I k +1 ;
- considering the previous frame as the current frame for the next iteration, if any;
wherein, for the first iteration, the current frame is the reference frame I n and the previous frame is a frame preceding the reference frame I n in the sequence;
and wherein said matrix is determined by using information from at least one of a system model between the previous frame and the current frame, and of observations of the function of positions of the set of pixels in at least the current frame. |