Content recognition for targeting video advertisements
First Claim
1. A computer-implemented method, the method comprising:
- performing object recognition on visual content of a video to identify one or more first time positions in the video where a respective recognized object appears in the visual content;
performing speech recognition on the video to identify one or more second time positions in the video where respective recognized speech occurs;
determining, by one or more processors a respective confidence level for one or more pairs of corresponding first and second time positions wherein the respective confidence level is based on a determination of whether the recognized speech of the second time position of the pair describes the respective recognized object of the first time position of the pair;
associating a respective advertising targeting criterion with one or more of the one or more first time positions based on the respective confidence level and the respective recognized object of the one or more first time position; and
providing an advertisement and the video based on, at least, the respective advertising targeting criterion.
2 Assignments
0 Petitions
Accused Products
Abstract
Methods, systems, and apparatus, including computer program products, for providing advertisements. A plurality of advertisement targeting criteria is determined from a video stream or file. A first advertisement targeting criterion of the advertisement targeting criteria is associated with a first time position in the video. A second advertisement targeting criterion of the advertisement targeting criteria is associated with a second time position in the video. One or more first advertisements are provided, for presentation at or after the first time position in the video, based on the first advertisement targeting criterion. One or more second advertisements are provided, for presentation at or after the second time position in the video, based on the second advertisement targeting criterion.
158 Citations
18 Claims
-
1. A computer-implemented method, the method comprising:
-
performing object recognition on visual content of a video to identify one or more first time positions in the video where a respective recognized object appears in the visual content; performing speech recognition on the video to identify one or more second time positions in the video where respective recognized speech occurs; determining, by one or more processors a respective confidence level for one or more pairs of corresponding first and second time positions wherein the respective confidence level is based on a determination of whether the recognized speech of the second time position of the pair describes the respective recognized object of the first time position of the pair; associating a respective advertising targeting criterion with one or more of the one or more first time positions based on the respective confidence level and the respective recognized object of the one or more first time position; and providing an advertisement and the video based on, at least, the respective advertising targeting criterion. - View Dependent Claims (2, 3, 4, 5, 6)
-
-
7. A system comprising:
-
a non-transitory computer-readable medium having instructions stored thereon; and data processing apparatus programmed to; perform object recognition on visual content of a video to identify one or more first time positions in the video where a respective recognized object appears in the visual content; perform speech recognition on the video to identify one or more second time positions in the video where respective recognized speech occurs; determine a respective confidence level for one or more pairs of corresponding first and second time positions wherein the respective confidence level is based on a determination of whether the recognized speech of the second time position of the pair describes the respective recognized object of the first time position of the pair; associate a respective advertising targeting criterion with one or more of the one or more first time positions based on the respective confidence level and the respective recognized object of the one or more first time position; and provide an advertisement and the video based on, at least, the respective advertising targeting criterion. - View Dependent Claims (8, 9, 10, 11, 12)
-
-
13. A storage device having instructions stored thereon that, when executed by data processing apparatus, are operable to cause the data processing apparatus to perform operations comprising:
-
performing object recognition on visual content of a video to identify one or more first time positions in the video where a respective recognized object appears in the visual content; performing speech recognition on the video to identify one or more second time positions in the video where respective recognized speech occurs; determining a respective confidence level for one or more pairs of corresponding first and second time positions wherein the respective confidence level is based on a determination of whether the recognized speech of the second time position of the pair describes the respective recognized object of the first time position of the pair; associating a respective advertising targeting criterion with one or more of the one or more first time positions based on the respective confidence level and the respective recognized object of the one or more first time position; and providing an advertisement and the video based on, at least, the respective advertising targeting criterion. - View Dependent Claims (14, 15, 16, 17, 18)
-
Specification