NEW TOOL FOR APPROACHING E-LEARNING:

VIDEORDER™

Videorder™ Voice-based Speech Recognition and Language Processing Search

Technology with Finder™ Engine

Ferenc Kiss, Lia Bassa

Department of Information and Knowledge Management

Budapest University of Technology and Economics

Foundation for Informations Society

Viktor Justin

Your Service Media Communication Agency

Keywords: Speech recognition, voice based search technology, language processing, teaching application for E-

learning.

Abstract: Videorder™ Voice-based Speech Recognition and Language Processing search technology with Finder™

engine from @Your Service Media Communication Agency offers a market ready solution today, for

speech recognized searching at any kind of film material on the basis of audible information inside, like

words, sentences. All this independent of the language involved! The main strength of the system is

Language Processing and Speech Recognition in one application package. Videorder™ allows you to search

any video or audio clips that are relevant to your query. The method is a good example how the basic aim of

ICEIS can be implemented: bringing together the achievements of a research team with trainers of

information and knowledge management as well as with a Foundation for Information Society becoming the

practitioner of the program.

1 FILM AS SOURCE OF

INFORMATION

The film is the most concise form for

communicating information. In this case it also

means various sorts of audiovisual recordings like

lectures, presentations, illustrations to be

communicated, real life or fictional recordings. The

film aims at creating a new ‘frame of reference’ in

order to enhance the full participation and

collaboration between media researchers, education

authorities, media professional bodies, social and

political institutions and – of course – the users. In

order to encounter this tendency, citizens –

especially young people – should start exploring the

real possibilities of enabling themselves and the

future generations to give a critical and creative

answer to value-loaded image building. One of the

ways this can be achieved is the new medium: the

Internet. We strongly believe that future citizens,

individually or collectively, will use this new

communication medium to distribute their own

material and offer alternatives to corporate

‘mainstream’ images.

As a Hungarian writer said in the early 1920ies

about movie:

“Eye is like a faithful herald: as soon as it gets to

know something, it informs his master about it in a

short and succinct sentence. If the amount of the

new knowledge is so much that he cannot make a

full stop in his breathtaking report, the whole

information becomes a single stream as it happens to

a chain rushing in front of our eyes, we see it as a

single line. The chain of the subsequent sentences

turn into one straight line: movement. Eye handles

impressions as Melanesian indigenous do figures:

what is more than five, is all the same, it is too

314

Kiss F., Bassa L. and Justin V. (2007).

NEW TOOL FOR APPROACHING E-LEARNING: VIDEORDERâ

c - Videorderâ

c Voice-based Speech Recognition and Language Processing

Search Technology with Finderâ

c Engine.

In Proceedings of the Ninth International Conference on Enterprise Information Systems - HCI, pages 314-318

DOI: 10.5220/0002389003140318

 SciTePress

much. All the movements happening within one

second in reality are the consequences of infinite

number of functions. For our eyes, this infinite

number means the same as six subsequent pictures

coming from Edison’s projector: six equals infinity.”

The information provision by the means of recorded

audio and video messages has had a serious

disadvantage until now: its real time duration, its

linearity meaning that it cannot be broken into

searchable basic units independently of the sounding

or viewed effects. Therefore

Research and development of external tools to make

film information searchable is being carried on in

our days with great efforts.

The main idea is thus to develop new and innovative

analytical tools that will allow film, television or any

broadcasting consumers, teachers engaged in media

literacy programs and industry professionals to

better understand and fully appreciate the potential

in the media output. Our aim is to present a tool and

guideline for innovative possibilities - e-

participation channels - permitted by the use of the

new Information Communication Technologies

(ICTs).

If the above problem is solved, it also provides a

device for E-Learning projects. As the teaching

material can be handled by the learners, a tailor

made curriculum can be set up.

2 THE PROLIFERATION OF

MULTIMEDIA CONTENT

The volume of multimedia content is expanding at

an exponential rate. Terrestrial, Satellite and Cable

Television, Radio and streaming media (audio and

video) from the World Wide Web means that users

can choose from a wide variety of media. But the

growing mountain of content raises a critical

question: how do users find the information they

want, on demand? Until now this question was

unanswered and avoided.

3 VOICE-BASED SEARCH

TECHNOLOGY – SPEECH

RECOGNITION - OVERVIEW

Current speech recognition technology has high

word error rates for large vocabulary sizes. There is

very little repetition in queries, providing the small

amount of information that could be used to guide

the speech recognizer. In speech recognition

applications, the recognizer can use context, such as

a dialogue history, to set up certain expectations and

guide the recognition. Voice search queries lack

such context. Voice queries can be very short (on the

order of only a few words or single word), so there

is very little information in the utterance itself upon

which a voice recognition determination can be

made. The hunt for the right videoclip on the Web

today is almost entirely dependent on metadata, the

tags and descriptions that identify the who, what and

where. Many video site operators believe that

metadata is the best tool for searching and all we

may need. Nevertheless metadata needs to be

recognized first. Speech recognition is the field of

computer science dealing with designing computer

systems that can recognize spoken words. Note that

voice recognition implies only the computer taking

dictation, not that it understands what is being said.

Comprehending human languages falls under a

different field of computer science called natural

language processing.

A number of voice recognition systems are available

on the market. The most powerful can recognize

thousands of words. However, they generally require

an extended training session during which the

computer system becomes accustomed to a

particular voice and accent. Such systems are said to

be speaker dependent. Many systems also require

that the speaker speak slowly and distinctly and

separate each word with a short pause. These

systems are called discrete speech systems.

Recently, great strides have been made in

continuous speech systems - voice recognition

systems that allow you to speak naturally. There are

now several continuous-speech systems available for

personal computers. Because of their limitations and

high cost, voice recognition systems have

traditionally been used only in a few special

situations.

4 VIDEORDER™

Videorder™ Voice-based Speech Recognition and

Language Processing search technology with

Finder™ engine from @Your Service Media

Communication Agency offers a market ready

solution today, for speech recognized searching at

any kind of film material on the basis of audible

information inside, like words, sentences. All this

independent of the language involved!

A NEW TOOL FOR APPROACHING E-LEARNING: VIDEORDERtm - Videorder™ Voice-based Speech Recognition

and Language Processing Search Technology with Finder™ Engine

315

Figure 2: The process of the technology.

Figure 1: The system of Videorder.

ICEIS 2007 - International Conference on Enterprise Information Systems

316

The main strength of the system is Language

Processing and Speech Recognition in one

application package.

The advantages of this system are the following:

• Language independent speech recognition

• 100% error free queries

• Use with mobile services

• Very short search times (especially

compared to real-time-viewing)

• Vast searchable databases of any kind of

filmed material

• Educational use for corporate trainings or at

universities

• Educational use for distance learning

systems

• Logical add-on for internet based film-

sharing portals

• Wide development capacity

5 TECHNICAL FEATURES

• Language independent speech recognition

• 100% error free queries

• Xml based

• Integrated metadata GUI

(means: real time search parallel with watching

video stream, playing presentations, or any kind

of additional media file)

• Result exported to any kind of database.

Videorder™ combines the most advanced speech

recognition technology with intelligent text analysis

and synchronization technologies to deliver

unparalleled automation and retrieval of multimedia

content.

Once feeding the input video or audio stream

into Videorder™, it is using advanced image and

audio analysis techniques to extract information of a

video or audio file in real time and to create a rich

index about the video or audio content. This

complete, precise, time-stamped index provides fine-

grained access to the video content that you can use

for searching and efficiently locate a specific video

segment for playback. For indexing Videorder™

uses advanced indexing technology, so the user can

quickly locate specific segments within the video

content or audio clip. Videorder™ generates

metadata tracks to save information generated by the

media analysis process. The information in all

metadata tracks are time stamped and synchronized

with the associated digital video file. Videorder™

platform controls both the indexing and encoding

processes to ensure synchronization between the

metadata captured from the video asset and the

associated digital file.

6 ADVANCED VIDEO/ AUDIO

ANALYSIS

Videorder™ utilizes advanced speech recognition

techniques and synchronization technologies to

analyze and understand the actual content (spoken

words) of an audio/video file, delivering supreme

accuracy, and access to multimedia content in any

form. The output of these analytical sub-processes

are stored as further metadata tracks, alongside with

the digitally encoded content itself; not only does

Videorder™ know what was said, Videorder™

knows exactly when it was said.

7 CORE AUDIO TECHNOLOGY

Videorder™ uses audio analysis technology that

applies advanced methods to deal with all aspects of

processing digital audio signals of an audio or video

stream. In order to analyse the spoken words of an

audio or video stream, the audio analysis techniques

used by Videorder™ are based on neural network

technology that is able to provide a fast, accurate

and dynamic solution within variable and rapidly

changing acoustic environments.

Rather than just relying on the existing metadata for

the description of an audio or video clip,

Videorder™ actually provides the ability to retrieve

a wide range of multimedia content, based on the

spoken words that were actually said in the

television or radio clip. This opens a whole new

realm of possibilities for accessing multimedia

online or in a database.

8 NON-DICTATED SPEECH

Some information feeds, like e.g. news broadcasts

and radio, are often intrinsically difficult to

transcribe due to noisy conditions and less than

perfect articulation. Videorder™ having an

additional, special feature, the sophisticated “Human

Ears” technology enables it to filter out extraneous

noise, to compensate low volume levels predict

intended dialogue with high probability. Approaches

A NEW TOOL FOR APPROACHING E-LEARNING: VIDEORDERtm - Videorder™ Voice-based Speech Recognition

and Language Processing Search Technology with Finder™ Engine

317

to searching through information using simple

keyword query mechanisms fail because the context

provided is not enough. In such cases users are

flooded by hundreds or thousands results, all of

which match the query and many representing the

different contexts within which the query term or

terms can appear in. Videorder™ is unique in not

second-guessing a user's need but seamlessly

provides the automated guidance required to enrich

the overall experience. Once the video or audio

stream has been indexed, encoded and analyzed the

digital files are stored on the Videorder™ server

from where you can retrieve and play the digital files

through the Videorder™ web browser. The

Videorder™ video server provides fast storage and

retrieval for all digital files of the indexed

video/audio channel.

9 VIDEORDER™ SEARCH

Videorder™ allows you to search any video or audio

clips that are relevant to your query. The

Videorder™ query refinement feature allows

selecting the best results of a query in order to

produce even more relevant results. Sorting by

relevance or dating Videorder™ enables the user to

sort audio and video results by conceptual relevance

or date.

10 CONCLUSION

The next step is to prepare a pilot of the training

program where supposedly the participants will be

actively involved in building and maintaining

(moderating) their own transnational communities in

order to share knowledge and insights into political,

social and cultural collective decision-making for

the purpose of the introduction of this new technique

in training. In order to establish responsibility for

these new media of exchange, participants will

analyze existing virtual communities and the

discursive behavior observed in these fora. By

disseminating results and providing information for

people, we would like to promote awareness of the

E-Learning program actions and co-operation

between involved institutions, players and students,

the latter both interested in the further development

of the tool or in using it for their studies.

The above described method is a good example how

the basic aim of ICEIS can be implemented:

bringing together the achievements of the research

team of Videorder™ with trainers of information

and knowledge management as well as with the

Foundation for Information Society becoming the

practitioner of the program by buying, using and

disseminating it to its partners.

REFERENCES

Bindé, Jérôme, Towards knowledge societies. Paris,

UNESCO, 2005.

Illeris, Knud, Adult education and adult learning,

Malabar, FL, Krieger, 2004. 245 p. Bibl.: p.231-242

James, Kathryn; Nightingale, Christine, Discovering

potential: a practitioner's guide to supporting

improved self-esteem and well-being through adult

learning. Leicester, UK, NIACE, 2004. 85 p.,

Longworth, Norman, Enthusing and empowering learners.

Leicester, UK. pp.149-159. (In: Journal of adult and

continuing; vol.10, 2004, no.2).

Madhukar, Indira, Lifelong learning in the information

society. New Delhi, Authorspress, 2004.

Singh, Madhu, The social recognition of informal learning

in different settings and cultural contexts. Köln,

Germany, Böhlau, 2005.

Thinesse-Demel, Jutta, Museums, libraries and cultural

heritage: democratising culture, creating knowledge

and building bridges; report on the workshop held at

the CONFINTEA V Midterm Review Conference,

Bangkok, 2003. Hamburg, Germany, UNESCO

Institute for Education, 2005

Welton, Michael, Designing the just learning society: a

critical inquiry. Leicester, UK, NIACE, 2005. 250 p.

Youngman, Frank, Changer les choses: les programmes

de développement et la formation des éducateurs

d'adultes Bonn. pp.137-152. (In: Education des adultes

et développement; 65,2005). Bibl.: p.152.

Contribuer à un avenir plus viable: éducation de qualité,

compétences nécessaires dans la viecourante et

éducation du développement durable UNESCO Paris,

UNESCO, Division de la promotion d'une éducation

de qualité, 2005. 6 p.,

If any, the appendix should appear directly after the

references without numbering, and not on a new page.

ICEIS 2007 - International Conference on Enterprise Information Systems

318