Image cropping according to points of interest
View Patent ↗In one embodiment, a computing device determines one or more points of interest in an image. The computing device selects a viewport position including a pre-determined number of the points of interest. The computing device crops the image based at least in part on the viewport position.
1. A method comprising:
by a computing device, determining one or more points of interest in an image, wherein each of the points of interest corresponds to (1) a user represented by a user node in a social graph of a social-networking system or (2) a concept represented by a concept node in the social graph;
by the computing device, assigning a weight for each of the one or more points of interest in the image based on social-networking information accessed from the social graph of the social-networking system;
by the computing device, analyzing a plurality of viewport test positions, each comprising at least a pre-determined number of the weighted points of interest, wherein the analyzing comprises:
setting boundaries of the viewport test position wherein at least one of the boundaries coincides with a position of one of the weighted points of interest; and
identifying any weighted points of interest in the viewport when the viewport is located at the viewport test position;
by the computing device, selecting one of the analyzed viewport test positions based on a respective number of the identified weighted points of interest and their assigned weights, and further based on a padding amount in the viewport; and
by the computing device, cropping the image based at least in part on the selected viewport test position.
2. The method of claim 1 , wherein at least one of the points of interest comprises at least a portion of an image of a person.
3. The method of claim 1 , wherein the at least one of the boundaries is horizontal.
4. The method of claim 1 , wherein the at least one of the boundaries is vertical.
5. One or more computer-readable non-transitory storage media embodying software that is operable when executed to:
determine one or more points of interest in an image, wherein each of the points of interest corresponds to (1) a user represented by a user node in a social graph of a social-networking system or (2) a concept represented by a concept node in the social graph;
assign a weight for each of the one or more points of interest in the image based on social-networking information accessed from the social graph of the social-networking system;
analyze a plurality of viewport test positions, each comprising at least a pre-determined number of the weighted points of interest, wherein the analyzing comprises:
setting boundaries of the viewport test position wherein at least one of the boundaries coincides with a position of one of the weighted points of interest; and
identifying any weighted points of interest in the viewport when the viewport is located at the viewport test position;
select one of the analyzed viewport test positions based on a respective number of the identified weighted points of interest and their assigned weights, and further based on a padding amount in the viewport; and
crop the image based at least in part on the selected viewport test position.
6. The media of claim 5 , wherein at least one of the points of interest comprises at least a portion of an image of a person.
7. The media of claim 5 , wherein the at least one of the boundaries is horizontal.
8. The media of claim 5 , wherein the at least one of the boundaries is vertical.
9. A system comprising:
one or more processors; and
a memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to:
determine one or more points of interest in an image, wherein each of the points of interest corresponds to (1) a user represented by a user node in a social graph of a social-networking system or (2) a concept represented by a concept node in the social graph;
assign a weight for each of the one or more points of interest in the image based on social-networking information accessed from the social graph of the social-networking system;
analyze a plurality of viewport test positions, each comprising at least a pre-determined number of the weighted points of interest, wherein the analyzing comprises:
setting boundaries of the viewport test position wherein at least one of the boundaries coincides with a position of one of the weighted points of interest; and
identifying any weighted points of interest in the viewport when the viewport is located at the viewport test position;
select one of the analyzed viewport test positions based on a respective number of the identified weighted points of interest and their assigned weights, and further based on a padding amount in the viewport; and
crop the image based at least in part on the selected viewport test position.
10. The system of claim 9 , wherein at least one of the points of interest comprises at least a portion of an image of a person.
11. The system of claim 9 , wherein the at least one of the boundaries is horizontal.
12. The system of claim 9 , wherein the at least one of the boundaries is vertical.