# US 2014/0193035 A1 · Method and Device for Head Tracking and Computer-Readable Recording Medium > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2014/0193035 A1 - Document ID: US20140193035A1 - Status: Published application - Kind: A1 - Application number: 14/129,709 - Filed: 2013-02-21 - Published: 2014-07-10 - Earliest claimed: 2012-02-23 - Assignees: Intel Corporation - Applicants: Intel Corporation - Inventors: Woo Ju Ryu - Web page: https://patentgrep.com/patent/US20140193035A1 ## CPC - G06T 7/251: {involving models} - G06V 10/7557: {based on appearance, e.g. active appearance models [AAM]} - G06V 40/167: {using comparisons between temporally consecutive images} - G06V 40/171: {Local features and components; Facial parts (eye characteristics G06V40/18); Occluding parts, e.g. glasses; Geometrical relationships} - G06T 2207/20081: Training; Learning - G06T 2207/30201: Face - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract The present disclosure is directed to performing head tracking. The method comprises: (a) receiving an input of an image including a facial area; and (b) tracking a movement of the facial area. The step (b) comprises: (b-1) if a rotation angle of a facial area is within a predetermined angle range from a front side, searching for a location change of feature points within a facial area through a comparison with a template learned in advance; and (b-2) if a rotation angle of a facial area is beyond a predetermined angle range from a front side, searching for a location change of feature points within a facial area through a comparison with a facial area image frame previously inputted. ## Claims ### Claim 1 A method for performing head tracking, comprising: (a) receiving an input of an image including a facial area; and (b) tracking a movement of the facial area, wherein the step (b) comprises: (b-1) if a rotation angle of a facial area is within a predetermined angle range from a front side, searching for a location change of feature points within a facial area through a comparison with a template learned in advance; and (b-2) if a rotation angle of a facial area is beyond a predetermined angle range from a front side, searching for a location change of feature points within a facial area through a comparison with a facial area image frame previously inputted. ### Claim 2 (depends on 1) The method of claim 1, further comprising: initializing by detecting a facial area from an inputted image. ### Claim 3 (depends on 2) The method of claim 2, wherein the step of initializing by detecting the facial area from the inputted image comprises: detecting feature points within a facial area using FD, ED and MD techniques from an inputted 2D image; and estimating an initial 3D location of a facial area from a facial area on the 2D image. ### Claim 4 (depends on 3) The method of claim 3, wherein the step of estimating the initial 3D location of the facial area comprises: estimating width and height of a face by using a distance between both eyes, estimating a roll angle using an angle between a line connecting both eyes and a horizontal line, and thereby calculating a rotation angle of a face. ### Claim 5 (depends on 2) The method of claim 2, wherein, after the step of initializing, a step of searching for a location change of feature points within a facial area through a comparison with the template learned in advance is performed at least once. ### Claim 6 (depends on 2) The method of claim 2, wherein the step of initializing is performed as angles of yaw, pitch, and roll are respectively within a range of ±22.5° from a front side of a facial area. ### Claim 7 (depends on 1) The method of claim 1, wherein the step (b-1) is implemented using an active appearance model (AAM) algorithm. ### Claim 8 (depends on 1) The method of claim 1, wherein the step (b-2) estimates a 3D location of a facial area using a dynamic template technique. ### Claim 9 (depends on 1) The method of claim 1, wherein the steps of searching for the location change of the feature points within the facial area are conducted by obtaining 6 parameters of the feature points. ### Claim 10 (depends on 9) The method of claim 9, wherein the 6 parameters include a rotation angle with respect to X axis (pitching), a rotation angle with respect to Y axis (yawing), a rotation angle with respect to Z axis (rolling), a parallel movement distance with respect to X axis, a parallel movement distance with respect to Y axis, and a parallel movement distance with respect to Z axis. ### Claim 11 (depends on 1) The method of claim 1, wherein the feature points within the facial area include centers of both eyes and both end points of a mouth. ### Claim 12 (canceled) ### Claim 13 An apparatus for performing head tracking, comprising: an image input unit configured to receive an input of an image including a facial area; and a facial area tracking unit configured to track a movement of the facial area, wherein the facial area tracking unit is configured to perform: a location change search of feature points within a facial area through a comparison with a template learned in advance, if a rotation angle of a facial area is within a predetermined angle range from a front side; and a location change search of feature points within a facial area through a comparison with a facial area image frame previously inputted, if a rotation angle of a facial area is beyond a predetermined angle range from a front side. ### Claim 14 (depends on 13) The apparatus of claim 13, further comprising: a facial area detecting unit configured to initialize by detecting a facial area from an inputted image. ### Claim 15 (depends on 14) The apparatus of claim 14, wherein the facial area detecting unit being configured to further perform: a detection of feature points within a facial area using FD, ED, and MD techniques from an inputted 2D image; and an estimate of an initial 3D location of a facial area from a facial area on the 2D image. ### Claim 16 (depends on 15) The apparatus of claim 15, wherein the estimate of the initial 3D location of the facial area comprises: an estimate of a width and height of a face by using a distance between both eyes, an estimate of a roll angle using an angle between a line connecting both eyes and a horizontal line, and thereby calculate a rotation angle of a face. ### Claim 17 (depends on 14) The apparatus of claim 14, wherein, after the initialization is performed, a process of searching for a location change of feature points within a facial area through a comparison with the template learned in advance is performed at least once. ### Claim at position 18 18-19. (canceled) ### Claim 20 (depends on 13) The apparatus of claim 13, wherein the location change search of the feature points within the facial area through the comparison with the facial area image frame previously inputted estimates a 3D location of a facial area using a dynamic template technique. ### Claim 21 (depends on 13) The apparatus of claim 13, wherein the location change search of the feature points within the facial area are conducted by obtaining 6 parameters of the feature points. ### Claim 22 (depends on 21) The apparatus of claim 21, wherein the 6 parameters include a rotation angle with respect to X axis (pitching), a rotation angle with respect to Y axis (yawing), a rotation angle with respect to Z axis (rolling), a parallel movement distance with respect to X axis, a parallel movement distance with respect to Y axis, and a parallel movement distance with respect to Z axis. ### Claim 23 (depends on 13) The apparatus of claim 13, wherein the feature points include centers of both eyes and both end points of a mouth. ### Claim 24 (depends on 13) The apparatus of claim 13, wherein the predetermined angle is that angles of yaw, pitch, and roll are respectively ±5° from a front of a facial area. ### Claim 25 A computer-readable medium having instructions, which when executed by a processor, causes the processor to perform: (a) receiving an input of an image including a facial area; and (b) tracking a movement of the facial area, wherein the step (b) comprises: (b-1) if a rotation angle of a facial area is within a predetermined angle range from a front side, searching for a location change of feature points within a facial area through a comparison with a template learned in advance; and (b-2) if a rotation angle of a facial area is beyond a predetermined angle range from a front side, searching for a location change of feature points within a facial area through a comparison with a facial area image frame previously inputted. ### Claim 26 (depends on 25) A computer-readable medium of claim 25, further comprising: initializing by detecting a facial area from an inputted image. ### Claim 27 (depends on 26) A computer-readable medium of claim 26, wherein the step of initializing by detecting the facial area from the inputted image comprises: detecting feature points within a facial area using FD, ED and MD techniques from an inputted 2D image; and estimating an initial 3D location of a facial area from a facial area on the 2D image. ### Claim 28 (depends on 27) A computer-readable medium of claim 27, wherein the step of estimating the initial 3D location of the facial area comprises: estimating width and height of a face by using a distance between both eyes, estimating a roll angle using an angle between a line connecting both eyes and a horizontal line, and thereby calculating a rotation angle of a face.