Browse Prior Art Database

System and Method for Generating Contextually Relevant Preview of a Video

IP.com Disclosure Number: IPCOM000239719D
Publication Date: 2014-Nov-27
Document File: 2 page(s) / 27K

Publishing Venue

The IP.com Prior Art Database

Related People

Vijay Mahadevan: INVENTOR [+5]

Abstract

A method and system is disclosed for generating a preview of a video. The preview contains frames of the video relevant to a context. The context is identified based on analysis of information. The information may include, but not limited to, search query, title and description of an article associated with the video.

This text was extracted from a Microsoft Word document.
This is the abbreviated version, containing approximately 57% of the total text.

System and Method for Generating Contextually Relevant Preview of a Video

Abstract

A method and system is disclosed for generating a preview of a video.  The preview contains frames of the video relevant to a context. The context is identified based on analysis of information.  The information may include, but not limited to, search query, title and description of an article associated with the video.

Description

Disclosed is a method and system for generating preview of a video.  The preview contains frames of the video relevant to context.  The context is identified based on analysis of information.  The information may include, but not limited to, search query, title and description of an article associated with the video.

In accordance with the method and system disclosed herein, each video from a video library are processed to create one or more previews for each video.  Each preview of a video is contextually relevant to one or more entities featuring in the video. 

In order to generate one or more previews of a video, shot boundaries are identified for the video.  For each shot boundary and for a given entity featuring in the video, the system analyzes frames in a shot.  For every frame, a visual tracking model is executed to recognize a given entity.  An entity can be recognized using face recognition or object recognition technologies.  Once, an entity is identified in a frame, the visual tracking model is executed to track the entity till the entity disa...