NEURAL NETWORK REAL-TIME CAPTIONING STENOGRAPHIC UNIT

Award Information
Agency:
Department of Education
Branch
n/a
Amount:
$30,000.00
Award Year:
1989
Program:
SBIR
Phase:
Phase I
Contract:
n/a
Agency Tracking Number:
10943
Solicitation Year:
n/a
Solicitation Topic Code:
n/a
Solicitation Number:
n/a
Small Business Information
Netrologic
5080 Shoreham Place, Suite 201, San Diego, CA, 92122
Hubzone Owned:
N
Socially and Economically Disadvantaged:
N
Woman Owned:
N
Duns:
n/a
Principal Investigator:
James R Johnson
() -
Business Contact:
() -
Research Institution:
n/a
Abstract
IT IS PRESENTLY VERY DIFFICULT FOR A DEAF PERSON TO SERVE ONA JURY OR AS A LAWYER IN A COURT ROOM. THE ONLY MEANS OF PROVIDING THE INFORMATION ABOUT THE PROCEEDINGS IS WITH AN INTERPRETER CAPABLE OF USING SIGN LANGUAGE. THIS PROPOSAL SUGGESTS THE POSSIBILITY OF USING THE COURT RECORDER AS A MEANS PROVIDING REAL TIME CAPTIONING IN THE COURT ROOM. NEURAL NETWORKS PROVIDE THE KEY TECHNOLOGY TO MAKE THIS POSSIBLE. THEY ARE A LEARNING TYPE SYSTEM WHICH FORM THEIR OWN INTERNAL STRUCTURE BASED ON LEARNING FROM EXAMPLES. A NEURAL NETWORK WOULD BE TRAINED ON A COURT RECORDERS TRANSCRIPT TO PROVIDE FULL TEXT. EACH RECORDER HAS INDIVIDUAL SHORT CUTS THAT THEY USE, BUT A NEURAL NET CAN BE TRAINED TO EXPAND THE ABBREVIATED FORM TO THE FULL FORM FROM COMPLETED EXAMPLES. THE RESULT WOULD BE REAL TIME, FULL TEXT TRANSCRIPTION AVAILABLE FOR USE IN THE COURTROOM.

* information listed above is at the time of submission.

Agency Micro-sites

US Flag An Official Website of the United States Government