Browse Prior Art Database

Discover and Highlight Speaker’s Identity in Web Conference Tool Disclosure Number: IPCOM000180525D
Original Publication Date: 2009-Mar-11
Included in the Prior Art Database: 2009-Mar-11
Document File: 5 page(s) / 87K

Publishing Venue



This article resovle the problem of multipart web-conference speaker identification,to make it more accurate identify the spearker.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 49% of the total text.

Page 1 of 5

Discover and Highlight Speaker's Identity in Web Conference Tool

1. Background: What is the problem solved by your invention? Describe known solutions to this problem(if any). What are the drawbacks of such known solutions, or why is an additional solution required? Cite any relevant technical documents or references.

In distributed team environment, tele-conf is usually used for verbal discussions, and web-conf is usually used for screen sharing, Instant Message chatting among team members.

The understanding of a team member's identity plays a critical role for effective communication and discussion to leverage all the context information (title, role, organization, etc) associated with the personidentity.

In many cases, the conference attendees have not met each other before, and even they don't know each other. Many attendees have hard time tolink the speakers'voice with their identities,

which brings negative impact to the effectiveness of the conference activity.

Tele-conf and Web-conf are usually separate systemswithout connectivity
This is a major issue if the conf has many attendees in enterprise environment.

2. Summary of Invention: Briefly describe the coreidea of your invention (saving the details for questions #3 below). Describe the advantage(s) of using your invention instead of the known solutionsdescribed above.

Discover speaker's identity through voice.

Visualize identity information.


Add one server side component to capture voice information from the tele-conf system.

Leverage speaker segmentation and recognition engine supplemented with human verification/correction when necessary to identify speaker.

Visualize the speaker identity in IM buddy list embedded in web-conf.

List potential speakers according to the meeting invitee list if it's hard to identify the speaker.

Provide an interface for users to modify the speaker identity, and synchronize the modification to other users. Design principal:
Don't involve changes to tele-conf system to minimize acquisition cost of this feature
The Speaker Segmentation and Recognition is a logiccomponent,

which can be implemented by different algorithms ortools.

3. Description: Describe how your invention works,and how it could be implemented, using text, diagrams and flow charts as appropriate.


Page 2 of 5

Fig 1 UI mock-up

Fig 1 is a UI mock-up of this invention. A speaker icon at the left of a user in the buddy list of Instant Messaging shows who is speaking at present. The meeting chairman can move/add the speaker icon to the real current speaker if the automatic speaker recognition generates a wrong identity.


[This page contains 2 pictures or other non-text objects]

Page 3 of 5

Fig 2 System architecture

Fig 2 is the architecture of this invention. The Tele-Conf Dialer dials in the Tele-Conf System, captures the voice and sends it...