A framework is suggested to segment the foreground in image sequences using background subtraction based on the reconstructed background image for each frame. First, the consecutive frames are taken as inputs for a motion estimation algorithm to calculate the motion vector between frames. Second, an incomplete background image can be achieved by cutting moving parts from origin images. Then, all the incomplete background images from the same sequence can be used to model the background by probabilistic principle component analysis with missing data. The background image can be estimated for each frame. Finally, a simple background subtraction can segment the foreground. The effectiveness of our method is demonstrated in different illumination conditions and compared to the commonly used Gaussian mixture models method.