Method and apparatus for interpolating frame to video, and electronic device

An electronic device, comprising:

• at least one processor;
• and
• a memory, communicatively coupled to the at least one processor;
• wherein,
• the memory is configured to store instructions executable by the at least one processor, when the instructions are executed by the at least one processor, the at least one processor is configured to:
• obtain a first deep-level feature of a first frame and a second deep-level feature of a second frame, the first frame and the second frame are adjacent frames of a video;
• obtain forward optical flow information and inverse optical flow information between the first frame and the second frame based on the first deep-level feature and the second deep-level feature;
• generate an interpolated frame between the first frame and the second frame based on the forward optical flow information and the inverse optical flow information, and insert the interpolated frame between the first frame and the second frame;
• wherein the at least one processor is configured to obtain the forward optical flow information by performing forward optical flow learning from the first frame to the second frame on the first deep-level feature and the second deep-level feature;
• and obtain the inverse optical flow information by performing inverse optical flow learning from the second frame to the first frame on the first deep-level feature and the second deep-level feature;
• and
• wherein the at least one processor is further configured to:
• generate a third deep-level feature by splicing the first deep-level feature and the second deep-level feature, the third deep-level feature being used for performing the forward optical flow learning and the inverse optical flow learning.

A non-transitory computer-readable storage medium, having computer instructions stored thereon, wherein the computer instructions are configured to cause a computer to execute a method for interpolating a frame to a video, the method comprising:

• obtaining a first deep-level feature of a first frame and a second deep-level feature of a second frame, the first frame and the second frame are adjacent frames of a video;
• obtaining forward optical flow information and inverse optical flow information between the first frame and the second frame based on the first deep-level feature and the second deep-level feature;
• generating an interpolated frame between the first frame and the second frame based on the forward optical flow information and the inverse optical flow information, and inserting the interpolated frame between the first frame and the second frame;
• wherein, obtaining the forward optical flow information and the inverse optical flow information comprises:
• obtaining the forward optical flow information by performing forward optical flow learning from the first frame to the second frame on the first deep-level feature and the second deep-level feature;
• and
• obtaining the inverse optical flow information by performing inverse optical flow learning from the second frame to the first frame on the first deep-level feature and the second deep-level feature;
• and
• wherein the method further comprises:
• generating a third deep-level feature by splicing the first deep-level feature and the second deep-level feature, the third deep-level feature being used for performing the forward optical flow learning and the inverse optical flow learning.