Obtaining artist imagery from video content using facial recognition
An example method may include applying an automated face detection program implemented on a computing device to a plurality of training digital images associated with a particular TV program to identify a sub-plurality of the training digital images, each containing a single face of a particular person associated with the particular TV program. A set of feature vectors determined for the sub-plurality may be used to train a computational model of a face recognition program for recognizing the particular person in any given digital image. The face recognition program and the computational model may be applied to a runtime digital image associated with the particular TV program to recognize the particular person in the runtime digital image, together with geometric coordinates. The runtime digital image may be stored together with information identifying the particular person and corresponding geometric coordinates of the particular person in the runtime digital image.
1. A method comprising:
applying an automated face detection program implemented on a computing device to a first plurality of training digital images associated with a particular television (TV) program to identify a first sub-plurality of the training digital images, each of which contains a single face of a first particular person associated with the particular TV program;
based on a first set of feature vectors determined for the first sub-plurality of training digital images, training a first computational model of a computer-implemented face recognition program for recognizing the first particular person in any given digital image;
applying the face recognition program together with the first computational model to a runtime digital image associated with the particular TV program to recognize the first particular person in the runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the runtime digital image; and
storing, in non-transitory computer-readable memory, the runtime digital image together with information identifying the recognized first particular person and corresponding geometric coordinates of the recognized first particular person in the runtime digital image.
2. The method of claim 1 , further comprising determining the first set of feature vectors, wherein each feature vector of the first set of feature vectors corresponds to a respective one of the first sub-plurality of training digital images.
3. The method of claim 1 , further comprising storing the first computational model, together with an identifier of the particular TV program and an identifier of the first particular person, in a database of computational models of the face recognition program.
4. The method of claim 1 , wherein applying the face recognition program together with the first computational model to the runtime digital image associated with the particular TV program to recognize the first particular person in the runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the runtime digital image comprises:
determining a runtime feature vector corresponding to at least one of the one or more faces detected together with respective geometric coordinates, in the runtime digital image;
applying the face recognition program together with one or more stored computational models associated with the particular TV program, including the first computational model, to the runtime feature vector; and
determining that applying the face recognition program together with the first computational model to the runtime feature vector yields a probability that both exceeds a threshold and is greater than probabilities yielded from applying the face recognition program together with any of the other one or more stored computational models associated with the particular TV program.
5. The method of claim 1 , wherein the computer-implemented face recognition program is implemented on one of: the same computing device as the automated face detection program, or a different computing device than the automated face detection program.
6. The method of claim 1 , wherein the first particular person is a cast member of the particular TV program.
7. The method of claim 1 , further comprising:
applying the face recognition program together with the first computational model to an additional runtime digital image associated with the particular TV program to again recognize the first particular person in the additional runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the additional runtime digital image; and
storing, in the non-transitory computer-readable memory, the additional runtime digital image together with information identifying the again recognized first particular person and corresponding geometric coordinates of the again recognized first particular person in the additional runtime digital image.
8. The method of claim 1 , further comprising:
applying the automated face detection program to a third plurality of training digital images associated with a further particular TV program to identify a third sub-plurality of training digital images, each of which contains a single face of a third particular person associated with the further particular TV program;
based on a third set of feature vectors determined for the third sub-plurality of training digital images, training a third computational model of the face recognition program for recognizing the third particular person in any given digital image;
applying the face recognition program together with the third computational model to a further runtime digital image associated with the further particular TV program to recognize the third particular person in the further runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the further runtime digital image; and
storing, in the non-transitory computer-readable memory, the further runtime digital image together with information identifying the recognized third particular person and corresponding geometric coordinates of the recognized third particular person in the further runtime digital image.
9. The method of claim 1 , further comprising:
applying the automated face detection program to a second plurality of training digital images associated with the particular TV program to identify a second sub-plurality of the training digital images each of which contains a single face of a second particular person associated with the particular TV program, wherein the second particular person is different than the first particular person;
based on a second set of feature vectors determined for the second sub-plurality of training digital images, training a second computational model of the face recognition program for recognizing the second particular person in any given digital image;
applying the face recognition program together with the second computational model to a further runtime digital image associated with the particular TV program to recognize the second particular person in the further runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the further runtime digital image; and
storing, in the non-transitory computer-readable memory, the further runtime digital image together with information identifying the recognized second particular person and corresponding geometric coordinates of the recognized second particular person in the further runtime digital image.
10. The method of claim 9 , wherein the first plurality and second plurality of training digital images are at least partially overlapping.
11. The method of claim 9 , wherein the runtime digital image and the further runtime digital image are the same common runtime digital image,
and wherein both (i) the information identifying the recognized first particular person and corresponding geometric coordinates of the recognized first particular person in the runtime digital image, and (ii) the information identifying the recognized second particular person and corresponding geometric coordinates of the recognized second particular person in the further runtime digital image, are stored with the common runtime digital image.
12. A system comprising:
one or more processors; and
memory storing instructions that, when executed by the one or more processors, cause the system to carry out operations including:
applying an automated face detection algorithm to a first plurality of training digital images associated with a particular television (TV) program to identify a first sub-plurality of the training digital images, each of which contains a single face of a first particular person associated with the particular TV program;
based on a first set of feature vectors determined for the first sub-plurality of training digital images, training a first computational model of an automated face recognition algorithm for recognizing the first particular person in any given digital image;
applying the automated face recognition algorithm together with the first computational model to a runtime digital image associated with the particular TV program to recognize the first particular person in the runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the runtime digital image; and
storing, in non-transitory computer-readable memory, the runtime digital image together with information identifying the recognized first particular person and corresponding geometric coordinates of the recognized first particular person in the runtime digital image.
13. The system of 12 , wherein the system further comprises a database of computational models of the face recognition program,
and wherein the operations further include storing the first computational model, together with an identifier of the particular TV program and an identifier of the first particular person, in the database of computational models of the face recognition program.
14. The system of claim 12 , wherein the one or more processors are comprised in two or more computing devices of the system,
and wherein face recognition program is implemented on one of: the same computing device of the system as the automated face detection program, or a different computing device of the system than the automated face detection program.
15. The system of claim 12 , wherein the operations further include:
applying the face recognition program together with the first computational model to an additional runtime digital image associated with the particular TV program to again recognize the first particular person in the additional runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the additional runtime digital image; and
storing, in the non-transitory computer-readable memory, the additional runtime digital image together with information identifying the again recognized first particular person and corresponding geometric coordinates of the again recognized first particular person in the additional runtime digital image.
16. The system of claim 12 , wherein the operations further include:
applying the automated face detection program to a second plurality of training digital images associated with the particular TV program to identify a second sub-plurality of the training digital images each of which contains a single face of a second particular person associated with the particular TV program, wherein the second particular person is different than the first particular person;
based on a second set of feature vectors determined for the second sub-plurality of training digital images, training a second computational model of the face recognition program for recognizing the second particular person in any given digital image;
applying the face recognition program together with the second computational model to a further runtime digital image associated with the particular TV program to recognize the second particular person in the further runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the further runtime digital image; and
storing, in the non-transitory computer-readable memory, the further runtime digital image together with information identifying the recognized second particular person and corresponding geometric coordinates of the recognized second particular person in the further runtime digital image.
17. The system of claim 16 , wherein the runtime digital image and the further runtime digital image are the same common runtime digital image,
and wherein both (i) the information identifying the recognized first particular person and corresponding geometric coordinates of the recognized first particular person in the runtime digital image, and (ii) the information identifying the recognized second particular person and corresponding geometric coordinates of the recognized second particular person in the further runtime digital image, are stored with the common runtime digital image.
18. A non-transitory computer-readable medium having instructions stored thereon that, when executed by one or more processors of a system, cause the system to carry out operations including:
applying an automated face detection algorithm to a first plurality of training digital images associated with a particular television (TV) program to identify a first sub-plurality of the training digital images, each of which contains a single face of a first particular person associated with the particular TV program;
based on a first set of feature vectors determined for the first sub-plurality of training digital images, training a first computational model of an automated face recognition algorithm for recognizing the first particular person in any given digital image;
applying the automated face recognition algorithm together with the first computational model to a runtime digital image associated with the particular TV program to recognize the first particular person in the runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the runtime digital image; and
storing, in non-transitory computer-readable memory, the runtime digital image together with information identifying the recognized first particular person and corresponding geometric coordinates of the recognized first particular person in the runtime digital image.
19. The non-transitory computer-readable medium of claim 18 , wherein the operations further include:
applying the automated face detection program to a second plurality of training digital images associated with the particular TV program to identify a second sub-plurality of the training digital images each of which contains a single face of a second particular person associated with the particular TV program, wherein the second particular person is different than the first particular person;
based on a second set of feature vectors determined for the second sub-plurality of training digital images, training a second computational model of the face recognition program for recognizing the second particular person in any given digital image;
applying the face recognition program together with the second computational model to a further runtime digital image associated with the particular TV program to recognize the second particular person in the further runtime digital image from among one or more faces detected, together with respective geometric coordinates, in the further runtime digital image; and
storing, in the non-transitory computer-readable memory, the further runtime digital image together with information identifying the recognized second particular person and corresponding geometric coordinates of the recognized second particular person in the further runtime digital image.
20. The non-transitory computer-readable medium of claim 18 , wherein the first particular person is a cast member of the particular TV program.