Method and system for real-time control of three-dimensional models

A method for real-time control of a three-dimensional model, comprising:

• capturing a real-time video of a head and a face in an actual object;
• locating a face region by utilizing a low-resolution copy of frame of images in the real-time video;
• applying directly the face region on a corresponding copy of an adjacent frame of images;
• marking a face landmark in the face region of the frame of images or the corresponding copy;
• establishing a head orientation reference pattern, a face reference planar and a face reference pattern of the face reference planar by utilizing a position fixed landmark of a front view 2D face in an image, and forming a coordinate map relationship with a front view 3D head model;
• capturing rotation data of a head by measuring deformation of the head orientation reference pattern relative to the face reference pattern when the head of the adjacent frame of images is rotating;
• and
• forming a control instruction of an action expression of the head and the face in combination with a position change of a 2D face landmark of adjacent frames and the rotation data of the head.

A system for real-time control of a three-dimensional model, comprising:

• a video capturing device (10), configured to obtain a real-time video of an actual object;
• an image marking device (20), configured to mark an action of the actual object in an image of the real-time video;
• and
• an action instruction generating device (30), configured to form an action control instruction of a corresponding 3D model according to a change of the action that is marked.