An image frame conversion method and a video frame conversion method and apparatus are provided. The video frame conversion method includes: extracting a longest line segment of a 2D video frame; determining a first depth image of the 2D video frame according to a direction of the longest line segment; determining motion information of the 2D video frame according to a video frame adjacent to the 2D video frame; determining a second depth image of the 2D video frame according to the motion information; and generating, according to the first depth image, the second depth image, and the 2D video frame, a 3D video frame corresponding to the 2D video frame.