System and Method for Video Editing Based on Semantic Data
AssigneeCyberlink Corp.
InventorBorching Lin
Life of this patent
Estimating the term…
Abstract
Systems and methods for video editing are described. At least one embodiment includes a method for editing a video comprising: extracting semantic data from the video, searching the extracted semantic data to identify characteristics that match a pre-defined set of characteristics, identifying a pre-defined event based on the matched characteristics within the semantic data, assigning a matching degree value to the event, assigning a theme relevancy value to the event, and automatically editing the event based on the identified event, the assigned matching degree value, and the theme relevancy value.
Claims25
A system for editing a video comprising:
- a module configured to extract semantic data from the video;
- a module configured to search the extracted semantic data to identify characteristics that match a set of pre-defined characteristics;
- and
- a module configured to identify a pre-defined event based on the matched characteristics within the semantic data.
A method for editing a video comprising:
- extracting semantic data from the video;
- searching the extracted semantic data to identify characteristics that match a pre-defined set of characteristics;
- and
- identifying a pre-defined event based on the matched characteristics within the semantic data;
A computer readable medium having a computer program for editing a video comprising:
- logic configured to extract semantic data from the video;
- logic configured to search the extracted semantic data to identify characteristics that match a set of pre-defined characteristics;
- logic configured to identify a pre-defined event based on the matched characteristics;
- logic configured to assign a matching degree value to the event;
- logic configured to assign a theme relevancy value to the event;
- and
- logic configured to automatically edit the video based on the identified event, the assigned matching degree value, and the theme relevancy value.
Dependencies are inferred from the claim text.