Aiming at the problem of jitter and bleeding in the particular application of 2 D-based video face replacement

a non-edge segmentation algorithm of video face anti-jitter and bleeding is proposed. Active shape model（ ASM） feature points are located on source faces and target faces

and source face is aligned to the target face by the Procrustes alignment algorithm

and the non-optimal face non-edge segmentation contour is obtained by linear interpolation in the non-edge region of the face. Energy function that takes into account the time continuity of the non-edge segmentation contour of the front and back frames and the texture smoothness of the non-edge segmentation contour of the current frame is designed. Dynamic programming algorithm is used to make energy minimization

and to get the optimal face non-edge segmentation contour. The optimal face non-edge segmentation contours of the subsequent frames are obtained. Experimental results show that this algorithm can effectively prevent the phenomenon of jitter and bleeding after replacement of the face

and has certain application value.