Method and apparatus for detecting face image

An apparatus for detecting a face image, comprising:

• at least one processor;
• and
• a memory storing instructions, wherein the instructions when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising:
• performing facial keypoint detection on an image frame containing a face object;
• segmenting a center face area image and an entire face area image of a target face object from the image frame based on a facial keypoint detection result, the center face area image being a face image excluding a fringe face area, and the entire face area image being a face image containing the fringe face area;
• performing head pose estimation respectively based on the center face area image and the entire face area image;
• and
• generating a detection result on whether the target face object contained in the image frame is a fake face object based on a difference between a piece of first head pose information estimated based on the center face area image and a piece of second head pose information estimated based on the entire face area image,
• wherein the piece of first head pose information comprises a first head pose angle vector, and the piece of second head pose information comprises a second head pose angle vector;
• and
• the generating a detection result on whether the target face object contained in the image frame is a fake face object based on a difference between a piece of first head pose information estimated based on the center face area image and a piece of second head pose information estimated based on the entire face area image comprises:
• computing a deviation between the first head pose angle vector and the second head pose angle vector;
• and
• determining, in response to determining that the deviation between the first head pose angle vector and the second head pose angle vector is greater than a preset threshold, the target face object contained in the image frame being the fake face object.

A non-transitory computer readable medium, storing a computer program thereon, wherein the program, when executed by a processor, causes the processor to perform operations, the operations comprising:

• performing facial keypoint detection on an image frame containing a face object;
• segmenting a center face area image and an entire face area image of a target face object from the image frame based on a facial keypoint detection result, the center face area image being a face image excluding a fringe face area, and the entire face area image being a face image containing the fringe face area;
• performing head pose estimation respectively based on the center face area image and the entire face area image;
• and
• generating a detection result on whether the target face object contained in the image frame is a fake face object based on a difference between a piece of first head pose information estimated based on the center face area image and a piece of second head pose information estimated based on the entire face area image,
• wherein the piece of first head pose information comprises a first head pose angle vector, and the piece of second head pose information comprises a second head pose angle vector;
• and
• the generating a detection result on whether the target face object contained in the image frame is a fake face object based on a difference between a piece of first head pose information estimated based on the center face area image and a piece of second head pose information estimated based on the entire face area image comprises:
• computing a deviation between the first head pose angle vector and the second head pose angle vector;
• and
• determining, in response to determining that the deviation between the first head pose angle vector and the second head pose angle vector is greater than a preset threshold, the target face object contained in the image frame being the fake face object.