Browse Prior Art Database

Method and System for Advertising on Streaming Speech Converted To Text

IP.com Disclosure Number: IPCOM000200481D
Publication Date: 2010-Oct-15
Document File: 4 page(s) / 181K

Publishing Venue

The IP.com Prior Art Database

Related People

Greg Muchnik: INVENTOR [+3]

Abstract

A method and system is disclosed for advertising on streaming speech converted to text. The method converts unplanned dynamic contents in a video into text in real-time for real-time advertising.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 47% of the total text.

Method and System for Advertising on Streaming Speech Converted To Text

Abstract

A method and system is disclosed for advertising on streaming speech converted to text.  The method converts unplanned dynamic contents in a video into text in real-time for real-time advertising.

Description

Disclosed is a method and system for advertising on streaming speech converted to text.  The method converts unplanned dynamic contents in a video to text in real-time for real-time advertising.  The video may be for example, news channels or talks shows in a television or internet

For example, typically, television advertising is a static operation such that Ads are pre-determined for commercial breaks of the scheduled program.  However, many programs such as, talk shows or news channels contain dynamic content that are not planned.  This content can be converted to text and used for real-time advertising.  Ads can be displayed on a panel on the television (or webpage if being viewed online).  

The system as disclosed herein includes a viewing portal i.e., a television or a webpage and advertising platform.  The content broadcasted through the television or the webpage is converted to text.  The conversion of the content to the text may be performed in real-time using the known methods such as, a closed captioning technique.  The text is relayed to the web server that can fetch an Ad to be displayed on the viewing portal.  Upon receiving the text, the text is parsed to identify terms that have been bidded/purchased by advertisers.  While parsing the text, a context attribute that is passed to the advertising platform may be considered.  This context can be fetched from the TV listing to indicate what genre of show it is.  For example, knowing that a basketball game is going on, the term "court" (in context of basketball court) will have a different meaning than if it shows up during a show such as, Law & Order (in context of court of law).  In another example, a context associated with a video may be, major headlines associated with the video.  In this case, the major headlines may be “Major Oil Spill” or “World Cup”.

Thereafter, a relevant Ad is fetched from the Ad server.  For selecting the relevant Ads, an Ad selection processor receives certain data inputs.  The data inputs include metadata about a video i.e., a talk show, user information, raw speech converted to text, contextual look-up table of keywords.  The information pertaining to the video may include, video name, category, rating, etc.  Further, information associated with a user may include, age, gender, location, interests, history etc.  The data inputs may also include streaming text i.e., the text that flows when the video plays. 

The video metadata can be fetched from a TV guide interface or an online database of viewing information.  This database includes basic information about the viewing program such as what category of interest the viewing information i...