The IP.com Prior Art Database
English (United States)
2 pages / 24.8 KB
Page 01 of 2
Method of Dynamically Selecting Thumbnail Based on User 's Profile
Video thumbnail selection uses a variety of techniques, such as:
• User selection of a thumbnail
• A random keyframe, or keyframe at some duration into the video
• Measures of motion and spatial activity and the likeliness to include people
• Skin-color detection and face detection
• RMSE (Mean Square Error) Detecting, a combination of colors distribution
These techniques fail to consider the profile or preferences of the audience that will actually see the thumbnail.
For example, a user has been browsing a company's website. They visit the two different places on the site to learn about two different products. They later visit a page featuring product demo videos. The thumbnails for each video are agnostic to the viewer's profile and interests, missing a valuable opportunity to entice the viewer to watch a video featuring the company's product.
Prior art exists for generating video thumbnails based on properties and image contents of a video file. (1-3) However, the prior art fails to incorporate the target audience when selecting a representative thumbnail image.
Disclosed is a method of selecting a thumbnail image to represent a video based on the viewer's profile, interests, or preferences.
In the preferred embodiment:
1. A video is posted to an online website
2. A viewer visits the website where the video is posted
3. The hosting website selects the ideal thumbnail of the video based on the visitor's profile
4. Because the video's thumbnail is selected based on the visitor's interests, the visitor is enticed to watch the video
The process for implementing the invention in the preferred embodiment follows:
1. A video file is tagged with keywords that represent image content at v...