Extracting multiple facial photos from a video clip
View Patent ↗Extracting an optimal subset of facial photographs includes obtaining an initial set of facial photographs, removing from the initial set photographs any photographs that are of unacceptable quality, grouping a remaining set of photographs according to view angle, removing from the remaining set of photographs any photographs having an undesirable facial expression to provide a limited set of representative facial photographs, and, selecting, from the limited set of facial photographs, an optimal subset of facial photographs. Obtaining the initial set of photographs may include using a video camera while diversifying view angles and controlling recording quality. Obtaining the initial set of photographs may include obtaining a series of still images. The still images may be self-recorded by a person with a smartphone front-facing camera.
1. A method of extracting an optimal subset of facial photographs, comprising:
obtaining an initial set of facial photographs;
removing from the initial set of photographs any photographs that are of unacceptable quality;
grouping a remaining set of photographs according to lateral view angle of facial images on the photographs;
removing from the remaining set of photographs any photographs having an undesirable facial expression to provide a limited set of representative facial photographs; and
selecting, from the limited set of facial photographs, an optimal subset of facial photographs.
2. A method, according to claim 1 , wherein obtaining the initial set of photographs includes using a video camera while diversifying view angles and controlling recording quality.
3. A method, according to claim 1 , wherein obtaining the initial set of photographs includes obtaining a series of still images.
4. A method, according to claim 3 , wherein the still images are self-recorded by a person with a smartphone front-facing camera.
5. A method, according to claim 1 , wherein grouping the remaining set of photographs includes creating groups of photographs selected from a plurality of the following: photographs of a front full-face view, photographs of a profile view, photographs of a ¾ view, and photographs of a ⅔ view.
6. A method, according to claim 1 , wherein removing photographs of unacceptable quality includes removing from the initial set of photographs any photographs having an incomplete facial view, any blurry photographs, and any photographs having an extreme angle leaving too little of a recognizable facial area.
7. A method, according to claim 1 , wherein removing from the remaining set of photographs any photographs having an undesirable facial expression includes removing photographs showing at least one of: fear, anger, disgust, and sadness.
8. A method, according to claim 7 , wherein facial expressions are automatically detected by software.
9. A method, according to claim 1 , wherein selecting, from the limited set of facial photographs, an optimal subset of facial photographs includes having a user review several suboptimal candidate sets of facial photographs and manually select the optimal set of photographs therefrom.
10. A method, according to claim 1 , wherein selecting, from the limited set of facial photographs, an optimal subset of facial photographs includes using closeness of the limited set of facial photographs to ideal prototypes for particular facial expressions and angles and diversity of facial expressions of the limited set of facial photographs in a two-component multi-dimensional space of facial expressions and view angles.
11. A method of editing an electronic contact, comprising:
modifying text fields of the electronic contact;
removing from an initial set of photographs any photographs that are of unacceptable quality;
grouping a remaining set of photographs according to lateral view angle of facial images on the photographs;
removing from the remaining set of photographs any photographs having an undesirable facial expression to provide a limited set of representative facial photographs;
selecting, from the limited set of facial photographs, an optimal subset of facial photographs; and
adding at least one photograph of the optimal subset of facial photographs to the electronic contact.
12. A method, according to claim 11 , wherein the initial set of photographs includes frames obtained by using a video camera while diversifying view angles and controlling recording quality.
13. A method, according to claim 11 , wherein the initial set of photographs includes a series of still images.
14. A method, according to claim 13 , wherein the still images are self-recorded by a person with a smartphone front-facing camera.
15. A method, according to claim 11 , wherein grouping the remaining set of photographs includes creating groups of photographs selected from a plurality of the following: photographs of a front full-face view, photographs of a profile view, photographs of a ¾ view, and photographs of a ⅔ view.
16. A method, according to claim 11 , wherein removing photographs of unacceptable quality includes removing from the initial set of photographs any photographs having an incomplete facial view, any blurry photographs, and any photographs having an extreme angle leaving too little of a recognizable facial area.
17. A method, according to claim 11 , wherein removing from the remaining set of photographs any photographs having an undesirable facial expression includes removing photographs showing at least one of: fear, anger, disgust, and sadness.
18. A method, according to claim 17 , wherein facial expressions are automatically detected by software.
19. A method, according to claim 11 , wherein selecting, from the limited set of facial photographs, an optimal subset of facial photographs includes having a user review several suboptimal candidate sets of facial photographs and manually select the optimal set of photographs therefrom.
20. A method, according to claim 11 , wherein selecting, from the limited set of facial photographs, an optimal subset of facial photographs includes using closeness of the limited set of facial photographs to ideal prototypes for particular facial expressions and angles and diversity of facial expressions of the limited set of facial photographs in a two-component multi-dimensional space of facial expressions and view angles.
21. A non-transitory computer-readable medium containing computer software that extracts an optimal subset of facial photographs, the software comprising:
executable code that removes from an initial set photographs any photographs that are of unacceptable quality;
executable code that groups a remaining set of photographs according to lateral view angle of facial images on the photographs;
executable code that removes from the remaining set of photographs any photographs having an undesirable facial expression to provide a limited set of representative facial photographs; and
executable code that selects, from the limited set of facial photographs, an optimal subset of facial photographs.
22. A non-transitory computer-readable medium, according to claim 21 , wherein the initial set of photographs includes frames obtained using a video camera while diversifying view angles and controlling recording quality.
23. A non-transitory computer-readable medium, according to claim 21 , wherein the initial set of photographs includes a series of still images.
24. A non-transitory computer-readable medium, according to claim 23 , wherein the still images are self-recorded by a person with a smartphone front-facing camera.
25. A non-transitory computer-readable medium, according to claim 21 , wherein executable code that groups the remaining set of photographs creates groups of photographs selected from a plurality of the following: photographs of a front full-face view, photographs of a profile view, photographs of a ¾ view, and photographs of a ⅔ view.
26. A non-transitory computer-readable medium, according to claim 21 , wherein executable code that removes photographs of unacceptable quality removes from the initial of set photographs any photographs having an incomplete facial view, any blurry photographs, and any photographs having an extreme angle leaving too little of a recognizable facial area.
27. A non-transitory computer-readable medium, according to claim 21 , wherein executable code that removes from the remaining set of photographs any photographs having an undesirable facial expression removes photographs showing at least one of: fear, anger, disgust, and sadness.
28. A non-transitory computer-readable medium, according to claim 21 , wherein executable code that selects, from the limited set of facial photographs, an optimal subset of facial photographs uses closeness of the limited set of facial photographs to ideal prototypes for particular facial expressions and angles and diversity of facial expressions of the limited set of facial photographs in a two-component multi-dimensional space of facial expressions and view angles.
29. A non-transitory computer-readable medium containing software that edits an electronic contact, the software comprising:
executable code that modifies text fields of the electronic contact;
executable code that removes from an initial set of photographs any photographs that are of unacceptable quality;
executable code that groups a remaining set of photographs according to lateral view angle of facial images on the photographs;
executable code that removes from the remaining set of photographs any photographs having an undesirable facial expression to provide a limited set of representative facial photographs;
executable code that selects, from the limited set of facial photographs, an optimal subset of facial photographs; and
executable code that adds at least one photograph of the optimal subset of facial photographs to the electronic contact.
30. A non-transitory computer-readable medium, according to claim 29 , wherein the initial set of photographs includes frames obtained using a video camera while diversifying view angles and controlling recording quality.
31. A non-transitory computer-readable medium, according to claim 29 , wherein the initial set of photographs includes a series of still images.
32. A non-transitory computer-readable medium, according to claim 31 , wherein the still images are self-recorded by a person with a smartphone front-facing camera.
33. A non-transitory computer-readable medium, according to claim 29 , wherein executable code that groups the remaining set of photographs creates groups of photographs selected from a plurality of the following: photographs of a front full-face view, photographs of a profile view, photographs of a ¾ view, and photographs of a ⅔ view.
34. A non-transitory computer-readable medium, according to claim 29 , wherein executable code that removes photographs of unacceptable quality removes from the initial of set photographs any photographs having an incomplete facial view, any blurry photographs, and any photographs having an extreme angle leaving too little of a recognizable facial area.
35. A non-transitory computer-readable medium, according to claim 29 , wherein executable code that removes from the remaining set of photographs any photographs having an undesirable facial expression removes photographs showing at least one of fear, anger, disgust, and sadness.
36. A non-transitory computer-readable medium, according to claim 29 , wherein executable code that selects, from the limited set of facial photographs, an optimal subset of facial photographs uses closeness of the limited set of facial photographs to ideal prototypes for particular facial expressions and angles and diversity of facial expressions of the limited set of facial photographs in a two-component multi-dimensional space of facial expressions and view angles.