Back to results
US 9,003,289 B2Granted
Automatic detection of suggested video edits
AssigneeGoogle Inc.
InventorsJohn Gregg, William Brendel, Thomas Benton Bridgwater, Jason Matthew Toff
Life of this patent
Estimating the term…
Abstract
Systems and methods are provided herein relating to video classification. A trained classifier can analyze a video for suggested edits. A plurality of features of the video can be analyzed based that determines whether the video is a good candidate for various type of editing. The suggested edits can be performed automatically or with the authorization of a user who submitted the video. The user can review the edited video and either approve or reject the edits. Using an automated process to suggest and perform video edits can improve the quality of videos within a video data store.
Claims19
A system comprising:
- one or more computer processors;
- and
- a non-transitory memory that stores instructions, which when executed by the one or more computer processors, cause the one or more computer processors to perform operations, comprising:
- receiving a video;
- extracting one or more features from the video;
- determining video feature information for an extracted feature of the one or more features, wherein the determining is based on comparing the extracted feature to a first set of candidate features and a second set of candidate features associated with the extracted feature;
- training a classifier on the one or more features to identify whether the video is the first type of candidate for editing or the second type of candidate for editing, wherein the training comprises adding the extracted feature to the first set of candidate features or adding the extracted feature to the second set of candidate features based on the determining;
- determining whether the video is associated with a defined type of candidacy for editing based on the one or more features and training data, wherein based on a determination that the video is associated with the defined type of candidacy for editing, the classifier generates at least one suggested edit to perform on the video;
- generating an edited video by editing the video based on the at least one suggested edit;
- and
- sending for display, by a server, textual information descriptive of the at least one suggested edit for viewing by a user.
8 more elements
A method, comprising:
- employing a processor to execute computer executable instructions stored in a memory to perform operations comprising:
- extracting one or more features from a video;
- retrieving, from a data store, for an extracted feature of one or more extracted features, a first set of candidate features and a second set of candidate features;
- determining a video feature rating for the extracted feature, wherein the determining is based on comparing the extracted feature to the first set of candidate features and the second set of candidate features associated with the extracted feature;
- and
- training a classifier on the one or more extracted features to identify whether the video is a first type of candidate for editing or a second type of candidate for editing, wherein the training comprises adding the extracted feature to the first set of candidate features or adding the extracted feature to the second set of candidate features based on the determining.
3 more elements
A method, comprising:
- employing a processor to execute computer executable instructions stored in a memory to perform operations comprising:
- receiving a video from a user;
- extracting one or more features of the video;
- using a classifier to determine one or more suggested edits to the video based on the one or more features and a set of training data;
- after displaying the one or more suggested edits for viewing by the user, wherein the one or more suggested edits comprise textual information describing the one or more suggested edits, receiving authorization from the user to perform at least one of the one or more suggested edits;
- editing the video based on the authorized at least one of the one or more suggested edits to generate an edited video;
- determining a video feature rating for an extracted feature of the one or more features, wherein the determining is based on comparing the extracted feature to a first set of candidate features and a second set of candidate features associated with the extracted feature;
- and
- training the classifier on the one or more extracted features to identify whether the video is a first type of candidate for editing or a second type of candidate for editing, wherein the training comprises adding the extracted feature to the first set of candidate features or adding the extracted feature to the second set of candidate features based on the determining.
6 more elements
Dependencies are inferred from the claim text.