# US 2018/0322862 A1 · Method and apparatus to synthesize voice based on facial structures > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2018/0322862 A1 - Document ID: US20180322862A1 - Status: Published application - Kind: A1 - Application number: 16/039,053 - Filed: 2018-07-18 - Published: 2018-11-08 - Earliest claimed: 2014-09-25 - Assignees: Intel Corporation - Applicants: Intel Corporation - Inventors: Shamim Begum; Alexander A. Oganezov - Web page: https://patentgrep.com/patent/US20180322862A1 ## CPC - G10L 13/027: Concept to speech synthesisers; Generation of natural phrases from machine-based concepts (generation of parameters for speech synthesis out of text G10L13/08) - G06V 40/176: {Dynamic expression} - G10L 13/033: Voice editing, e.g. manipulating the voice of the synthesiser - G10L 13/047: Architecture of speech synthesisers - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract A method for establishing an articulatory speech synthesis model of a person's voice includes acquiring image data representing a visage of a person, in which the visage includes facial characteristics defining exteriorly visible articulatory speech synthesis model parameters of the person's voice; selecting a predefined articulatory speech synthesis model from among stores of predefined models, the selection based at least in part on one or both of the facial characteristics or the exteriorly visible articulatory speech synthesis model parameters; and associating at least a portion of the selected predefined articulatory speech synthesis model with the articulatory speech synthesis model of the person's voice. ## Claims ### Claim 1 A method for establishing an articulatory speech synthesis model of a person's voice, the method comprising: acquiring image data representing a visage of a person, in which the visage includes facial characteristics defining exteriorly visible articulatory speech synthesis model parameters of the person's voice; selecting a predefined articulatory speech synthesis model from among stores of predefined models, the selection based at least in part on one or both of the facial characteristics or the exteriorly visible articulatory speech synthesis model parameters; and associating at least a portion of the selected predefined articulatory speech synthesis model with the articulatory speech synthesis model of the person's voice. ### Claim 2 (depends on 1) The method of claim 1, in which the selection is based on a measure of a face-matching correlation between the facial characteristics of the visage of the person and facial characteristics defining visible articulatory speech synthesis model parameters of the predefined models. ### Claim 3 (depends on 2) The method of claim 2, in which the measure of face-matching correlation is derived using a hidden Markovian model. ### Claim 4 (depends on 1) The method of claim 1, in which the acquiring comprises capturing an image with an imager in a user equipment device. ### Claim 5 (depends on 1) The method of claim 1, in which the selecting comprises comparing the one or both of the facial characteristics or the exteriorly visible articulatory speech synthesis model parameters to those of the predefined models. ### Claim 6 (depends on 1) The method of claim 1, in which the selecting comprises communicating the image data from a user equipment device to a server for initiating a comparison of the one or both of the facial characteristics or the exteriorly visible articulatory speech synthesis model parameters to those of the predefined models. ### Claim 7 (depends on 1) The method of claim 1, in which the associating of the at least the portion of the selected predefined articulatory speech synthesis model comprises synthesizing speech using model parameters of the selected predefined articulatory speech synthesis model. ### Claim 8 (depends on 1) The method of claim 1, in which each one of the predefined models includes exteriorly visible articulatory speech synthesis model parameters defined by facial characteristics of a corresponding modeled voice and interiorly concealed articulatory speech synthesis model parameters defined by internal vocal tract characteristics of the corresponding modeled voice. ### Claim 9 (depends on 8) The method of claim 8, in which the associating of the at least the portion of the selected predefined articulatory speech synthesis model comprises synthesizing speech based on the exteriorly visible articulatory speech synthesis model parameters of the person's voice and based on the interiorly concealed articulatory speech synthesis model parameters of the selected predefined articulatory speech synthesis model. ### Claim 10 (depends on 8) The method of claim 8, in which the selection is based on a measure of correlation between the exteriorly visible articulatory speech synthesis model parameters of the person's voice and the exteriorly visible articulatory speech synthesis model parameters of the selected predefined articulatory speech synthesis model. ### Claim 11 (depends on 8) The method of claim 8, in which the associating of the at least the portion of the selected predefined articulatory speech synthesis model comprises communicating to a user equipment device the interiorly concealed articulatory speech synthesis model parameters of the selected predefined articulatory speech synthesis model for use in conjunction with exteriorly visible articulatory speech synthesis model parameters of the person's voice. ### Claim 12 (depends on 1) The method of claim 1, in which the acquiring of the image data comprises receiving an image file communicated over a network from a user equipment device. ### Claim 13 (depends on 12) The method of claim 12, in which the selecting comprises comparing of the one or both of the facial characteristics or the exteriorly visible articulatory speech synthesis model parameters to those of the predefined models. ### Claim 14 (depends on 1) The method of claim 1, in which the associating of the at least the portion of the selected predefined articulatory speech synthesis model comprises communicating to a user equipment device interiorly concealed articulatory speech synthesis model parameters of the selected predefined articulatory speech synthesis model. ### Claim 15 (depends on 14) The method of claim 14, further comprising combining the exteriorly visible articulatory speech synthesis model parameters of the person's voice with the interiorly concealed articulatory speech synthesis model parameters. ### Claim 16 (depends on 11) The method of claim 11, in which the stores of predefined models comprise a database, and the method further comprising: receiving a submission of a predefined model, the submission including a voice sample recording of a known text input; and determining whether a synthesized speech output produced using the predefined model matches the voice sample recording. ### Claim 17 (depends on 16) The method of claim 16, in which the stores of predefined models comprise a database, and the method further comprising rejecting the submission in response to the synthesized speech output insufficiently matching the voice sample recording. ### Claim 18 (depends on 1) At least one computer readable storage medium, having instructions stored thereon, that when executed on a device, configure the device to perform the method of claim 1.