Face detecting and tracking method, method for controlling rotation of robot head and robot

A method for controlling rotation of a robot head, comprising:

• determining a tracking box of a first image using a face detecting and tracking method, wherein the face detecting and tracking method comprises:
• acquiring an image and performing a face detection to determine one or more face boxes of each acquired image;
• performing a detection error removing process to the image of current frame, and saving one of the one or more face boxes of the processed image of current frame as a first face box;
• and
• determining a tracking box of the image of current frame based on the first face box and a tracked face box, and tracking the face in the tracking box of the image of current frame;
• determining if the tracking box of the first image is located at a central area of the first image;
• and
• if no, controlling a robot head to rotate until a tracking box of a second image captured after the first image is located at a central area of the second image;
• wherein the step of performing a detection error removing process to the image of current frame, and saving one of the one or more face boxes of the processed image of current frame as a first face box, comprises:
• determining that an image of previous frame, which the image of current frame is successive to, and the image of current frame comprise one or more face boxes;
• determining that the one or more face boxes of the image of current frame and the one or more face boxes of the image of previous frame are located at corresponding positions;
• and
• determining that the image of current frame does not comprise a falsely detected face box, and saving one of the one or more face boxes of the image of current frame as the first face box;
• wherein the step of determining if the one or more face boxes of the image of current frame and the one or more face boxes of the image of previous frame are located at corresponding positions, comprises:
• determining that the image of current frame and the image of previous frame are orientated in the same way in the same coordinate system, and that overlapping portions exist between the one or more face boxes of the image of current frame and the one or more face boxes of the image of previous frame;
• computing a first overlapping rate of an area of the overlapping portions to an area of the one or more face boxes of the image of current frame, and computing a second overlapping rate of an area of the overlapping portions to an area of the one or more face boxes of the image of previous frame;
• in response to the first overlapping rate and/or the second overlapping rate being greater than a predetermined value,
• determining that the one or more face boxes of the image of current frame and the one or more face boxes of the image of previous frame are located at corresponding positions.

A robot controlling rotation of a robot head using a method comprising:

• determining a tracking box of a first image using a face detecting and tracking method, wherein the face detecting and tracking method comprising:
• acquiring an image and performing a face detection to determine one or more face boxes of each acquired image;
• performing a detection error removing process to the image of current frame, and saving one of the one or more face boxes of the processed image of current frame as a first face box;
• and
• determining a tracking box of the image of current frame based on the first face box and a tracked face box, and tracking the face in the tracking box of the image of current frame;
• determining if the tracking box of the first image is located at a central area of the first image;
• and
• if no, controlling a robot head to rotate until a tracking box of a second image captured after the first image is located at a central area of the second image;
• wherein the step of determining the tracking box of the image of current frame based on the first face box and the tracked face box, and tracking the face in the tracking box of the image of current frame, comprises:
• comparing an area of the first face box with an area of the tracked face box;
• if the area of the first face box is greater than or equal to the area of the tracked face box,
• determining that the first face box is the tracking box of the image of current frame, and tracking the face in the tracking box of the image of current frame;
• and
• if the area of the first face box is less than the area of the tracked face box, determining that the tracked face box is the tracking box of the image of current frame, and tracking the face in the tracking box of the image of current frame.