Obtaining person-specific images in a public venue
View Patent ↗Collecting images of a patron in an entertainment venue is performed by facial recognition of the patron's face within the images, simplifying the storage and distribution of the images for a patron. In order to enhance the reliability of the facial recognition system, information about the patron that is not directly related to most facial recognition systems, including clothes, height, other associated people, use of glasses and jewelry, disposition of facial hair, and more, can be used. Some of the characteristics used can be specific to a particular date or event, and which will not be more generally characteristic of the patron. The facial recognition system can also be used to identify the patron requesting images to be collected. The present invention also relates to the use of facial recognition in sorting and collecting images from an electronically-stored image collection, enabling the easy retrieval of images that are related to a particular person or set of people.
1. A method of distributing to a user a specific electronic image comprising a representation of a specific person from among a collection of electronic images containing representations of many persons, the method comprising:
using a computer for performing the steps of: performing facial recognition analysis on each electronic image in the image collection by which a facial identifier is determined for each face represented within each electronic image;
associating two or more facial identifiers with the specific person, wherein a first facial identifier and a second facial identifier of the specific person are associated on the basis of the similarities of the first and second facial identifiers, in conjunction with additional information from the two electronic images from which the first and second identifiers are derived, wherein the additional information does not involve facial identifiers;
storing each electronic image in association with the facial identifier of each person represented within the electronic image;
obtaining a reference facial identifier from a source electronic image in which the specific person is represented;
selecting the specific electronic image for distribution on the basis of the similarity of the reference facial identifier to a facial identifier associated with the specific person, wherein the specific electronic image is associated with a facial identifier associated with the person.
2. The method of claim 1 , wherein the source electronic image is not within the collection.
3. The method of claim 1 , wherein the source electronic image is transmitted via a wide area digital communications network.
4. The method of claim 1 , wherein the source electronic image is within the collection, and wherein the user manually chooses the specific person representation from the source image following visual inspection of the image.
5. The method of claim 4 , wherein images from the collection are sorted for presentation to the user on the basis of the facial identifiers associated with the images.
6. The method of claim 5 , wherein images with the largest number of facial identifiers that the user has not previously been presented are presented to the user, wherein the user can efficiently locate the specific person in the collection of images by inspecting only a subset of the images.
7. The method of claim 1 , wherein the step of associating comprises a confidence level that is compared with a predetermined threshold, wherein a confidence level above the threshold indicates the association of a facial identifier to a person.
8. The method of claim 1 , wherein the additional information comprises the presence of a representation of the specific person in a first image and the presence of another representation of the specific person in an image consecutive with the first image, wherein the additional information increases the confidence level for associating the facial identifier for the face representative of the specific person with the specific person.
9. The method of claim 1 , wherein the additional information comprises the presence of a representation of the specific person in a first image and the presence of another representation of the specific person in a second image captured within a predetermined closeness in time with the first image, wherein the additional information increases the confidence level for associating the facial identifier for the face representative of the specific person with the specific person.
10. The method of claim 9 , wherein the additional information comprises the presence of a representation of the specific person in a first image and the presence of another representation of the specific person in a second image captured within a predetermined closeness in time and a predetermined proximity with the first image, wherein the additional information decreases the confidence level for associating the facial identifier for the face representative of the specific person with the specific person.
11. A method of associating a specific person with a specific electronic image within a collection of electronic images containing representations of many persons, the method comprising:
using a computer for performing the steps of: performing facial recognition analysis on each electronic image in the image collection by which a facial identifier is determined for each face represented within each electronic image;
calculating a confidence level for determining whether a given face representation in a first image is associated with the specific person, wherein the confidence level is computed from the similarity of the facial identifier of the given face representation and the facial identifier of a face representation contained within a second image from the collection, in conjunction with additional information from the first and second images;
associating the specific electronic image with the specific person if the confidence level exceeds a threshold.
12. The method of claim 11 , wherein the additional information that the second image is consecutive to the first image.
13. The method of claim 12 , wherein the additional information comprises that the second image was captured within a predetermined closeness in time to the first image.
14. The method of claim 13 , wherein additional information comprises that the second image was captured within a predetermined closeness in time and proximity to the first image.
15. The method of claim 12 , wherein the additional information comprises membership of the facial identifier within a first cluster of facial identifiers that has been formed using cluster analysis on facial identifiers from electronic images within the image collection.
16. The method of claim 15 , wherein the first cluster of facial identifiers is modified by joining the first cluster and a second cluster using manual input from the user following visual inspection of at least one image from each of the first and second clusters.
17. The method of claim 11 , wherein the additional information comprises that the second image was captured within a predetermined closeness in time and proximity to the first image.
18. The method of claim 11 , wherein the additional information comprises that the representation of a second person is present in both the first image and the second image.
19. A system for distributing to a user a specific electronic image comprising a representation of a specific person from among a collection of electronic images containing representations of many persons, comprising:
a facial recognizer that generates a facial identifier for each facial representation within each image;
an associator that associates each facial identifier with the corresponding person, wherein the associator determines a confidence level that is compared with a predetermined threshold, wherein a confidence level above the threshold indicates the association of a facial identifier to a person, and wherein the confidence level is computed from the similarity of facial identifiers and additional information that can be derived from two different images in the collection;
a storage device that stores each electronic image in association with the facial identifier of each person represented within the electronic image;
a reference facial identifier, which is obtained from a source electronic image in which the specific person is represented;
and a selector which receives the facial identifiers associated with each image from the storage device and which further receives the reference facial identifier and which determines which image to distribute, wherein the selector selects the specific electronic image for distribution on the basis of the similarity of the reference facial identifier to a facial identifier associated with the specific electronic image.
20. The system of claim 19 , wherein the interface and the retriever are connected to a wide area digital communications network for use by the user to specify the specific person.
21. The system of claim 19 , wherein the user supplies an electronic image comprising a representation of a face of the specific person.
22. The system of claim 19 , wherein the presence of a person in a first image and a second image captured within a predetermined closeness in time to the first image increases the confidence level for associating the facial identifier for the face representative of the first person with the first person.
23. The system of claim 19 , wherein the presence of a first and second person in a first image and the first and second person in a second image increases the confidence level for associating the facial identifier for the face representative of the first person with the first person.
24. The system of claim 19 , wherein the presence of a person in a first image and the image consecutive with the first image increases the confidence level for associating the facial identifier for the face representative of the first person with the first person.