Image-based face search
A search includes comparing a query image provided by a user to a plurality of stored images of faces stored in a stored image database, and determining a similarity of the query image to the plurality of stored images. One or more resultant images of faces, selected from among the stored images, are displayed to the user based on the determined similarity of the stored images to the query image provided by the user. The resultant images are displayed based at least in part on one or more facial features.
1. One or more computer-readable media storing instructions comprising
prompting a user to provide a query image of a face;
prompting the user to indicate a preference for one or more facial features of the query image of the face;
comparing the query image provided by the user to a plurality of stored images of faces stored in a stored image database wherein the comparing comprises computing a difference vector between the query image and each of the stored images;
determining a similarity of the query image provided by the user to the plurality of stored images of faces stored in the stored image database, based on the comparison and a mapping function of the stored images, wherein the mapping function is determined based on similarity scores assigned by one or more human assessors; and
displaying one or more resultant images of faces to a user, the resultant images being selected from among the stored images based on the determined similarity to the query image provided by the user, and based on the facial feature preferences indicated by the user.
2. The one or more computer-readable media of claim 1 , wherein prompting the user to indicate a preference for one or more facial features comprises prompting the user to assign a weighted preference value to the one or more facial features of the query image of the face.
3. The one or more computer-readable media of claim 1 , wherein prompting the user to indicate a preference for one or more facial features comprises prompting the user to indicate one or more of the facial features of the query image of the face as being more preferred than one or more other facial features.
4. The one or more computer-readable media of claim 1 , wherein the one or more facial features comprise eyes, mouth, nose, and shape of the face in the query image.
5. The one or more computer-readable media of claim 1 , wherein prompting the user to indicate a preference for one or more facial features comprises prompting the user to assign a weighted preference value for each of the eyes, mouth, nose, and shape of the face in the query image.
6. The one or more computer-readable media of claim 1 , wherein displaying the resultant images comprises clustering the resultant images into multiple groups based on facial features.
7. The one or more computer-readable media of claim 6 , wherein the groups into which the resultant images are clustered comprise a holistic group comprising resultant images that as a whole resemble the query image face, and at least one facial feature group comprising resultant images that have a specified facial feature that resembles a corresponding facial feature of the query image face.
8. The one or more computer-readable media of claim 6 , wherein the groups into which the resultant images are clustered comprise:
a holistic group comprising resultant images that as a whole resemble the query image face, an eye facial feature group comprising resultant images of faces having eyes that resemble eyes of the query image face,
a nose facial feature group comprising resultant images of faces having noses that resemble a nose of the query image face,
a mouth facial feature group comprising resultant images of faces having mouths that resemble the mouth of the query image face, and
a face shape facial feature group comprising resultant images of face shapes that resemble a face shape of the query image face.
9. The one or more computer-readable media of claim 1 , further comprising instructions for:
prompting the user to enter a textual description of one or more characteristics of a person; and
searching the database of stored images for stored images of faces corresponding to people having the characteristics specified by the user prior to comparing the query image to the plurality of stored images,
wherein the plurality of stored images of faces to which the query image is compared comprises only those stored images of faces corresponding to people having the characteristics specified by the user.
10. The one or more computer-readable media of claim 1 , further comprising instructions for:
prompting the user to enter profile information about the user including a picture of the user's face; and
storing the picture of user's face in the stored image database.
11. The one or more computer-readable media of claim 1 , wherein the query image may be provided by the user uploading a photograph, or by the user selecting an image from among the stored images.
12. An image-based online dating service at least partially implemented by the one or more computer-readable media of claim 1 .
13. A method comprising:
comparing a query image provided by a user to a plurality of stored images of faces stored in a stored image database wherein the comparing comprises computing a difference vector between the query image and each of the stored images; and
displaying one or more resultant images of faces to a user, the resultant images being selected from among the stored images based on a determined similarity to the query image provided by the user and a mapping function of the stored images, wherein the mapping function is determined based on similarity scores assigned by one or more assessors,
wherein the displayed resultant images are clustered into multiple groups based at least in part on one or more facial features.
14. The method of claim 13 , wherein the groups into which the resultant images are clustered comprise a holistic group comprising resultant images that as a whole resemble the query image face, and at least one facial feature group comprising resultant images that have a facial feature that resembles a corresponding facial feature of the query image face.
15. The method of claim 13 , wherein the groups into which the resultant images are clustered comprise:
a holistic group comprising resultant images that as a whole resemble the query image face, an eye facial feature group comprising resultant images of faces having eyes that resemble eyes of the query image face,
a nose facial feature group comprising resultant images of faces having noses that resemble a nose of the query image face,
a mouth facial feature group comprising resultant images of faces having mouths that resemble the mouth of the query image face, and
a face shape facial feature group comprising resultant images of face shapes that resemble a face shape of the query image face.
16. The method of claim 13 , further comprising prompting the user to indicate a preference for one or more facial features of the query image of the face, wherein the displayed resultant images are clustered into multiple groups based at least in part on the facial feature preferences indicated by the user.