SlideShare a Scribd company logo
BRANA (ብራና): APPLICATION OF AMHARIC SPEECH RCOGNITION SYSTEM
FOR
DICTATION IN JUDICIAL DOMAIN
Presenter: Bantegize Addis
Adviser: Solomon Teferra
June 05, 2015
What I’m going to talk about…
1. Introduction…
2. Related Works….
3. Speech database….
4. Architecture…
5. Implementation…
6. Our Results….
7. Conclusion & future works…
2
1. Introduction
General Background
 Automatic speech recognition gives us a new channel for
communication with computers.
 Speech technology is the technology of today and
tomorrow.
 It has practical implementations for both fun and serious
works.
 Mostly applied in command and control, data entry
and retrieval, and dictation functions.
3
Intro.(Cont.)
Statement of the Problem
 Dictation has been a common application area for ASRS for a
long period.
 Amharic is the second most-spoken Semitic language in the
world .
 It is the official working language of the FDRE.
 To the best of our knowledge, there is no attempted works
about application of speech recognition for dictation in
Amharic.
4
2. Related works
 Automatic speech recognition for Amharic
5
Author Type of
recognizer
Unit of
recognition
Recognition result
Solomon
Berhanu
Isolated Consonant-Vowel
(CV) Syllable
Speaker dependent : 87.68%
Speaker independent: 72.75%
Speaker not involved in the
training: 49.21%
Kinfe
Taddesse
Isolated Phoneme
Tri-phone
Consonant-Vowel
(CV) Syllable
Speaker independent tri-phone
(Test set I): 91.46 %
Speaker independent tri-phone
(Test set I): 77.87%
Related…(Cont.)
Author Type of
Recognizer
Unit of
Recognition
Recognition result
Solomon
Teferra
Continuous Tri-phone
Consonant-Vowel
(CV) Syllable
Speaker Independent tri-
phone: 91.31%
Speaker Independent CV: of
90.43%
6
Application of Amharic Speech Recognition
Martha developed Amharic speech input interface to command and control
Microsoft Word
3. Speech Corpus & annotation
 Two types of Speech Corpus
I. Spontaneous speech corpus
II. Read speech corpus
7
4. The Architecture of BRANA
8
5. Implementation of BRANA
Development Tools
 JDK jdk1.7.0_05 with Eclipse
 Sphinx-4
 Sphinx Trainer
 SRILM
 cmuclmtk-0.7-win32
9
Implementation…(Cont.)
User Interface Implementation
 The user interface of our system is responsible for:
 editing rtf text documents
 PPT(Push To Talk) functionality
 displaying the uttered word or sentence hypothesis to the
user
 Developed using Java
10
Implementation…(Cont.)
11
Push button to initiate recogntion
GUI snapshot of BRANA(ብራና)
Implementation…(Cont.)
Python program implementation
 Pronunciation Dictionary
Python script program is implemented for generating a
grapheme-based canonical pronunciation dictionary
12
6. Our results…
13
Recognition performance of spontaneous speech recognizer
(In a Batch Recognition Mode)
Acoustic
Model
Language
Model
Accuracy
(%)
WER (%)
AM-CD_8
LM-ABS 49.537 55.093
LM-GT 49.769 56.019
LM-MKN 50.463 54.861
LM-WB 49.769 55.093
Spontaneous
Speech Recognizer
result…(Cont.)
14
Recognition performance of continuous read speech recognizer
(In a Batch Recognition Mode)
Acoustic
Model
Language
Model
Accuracy
(%)
WER (%)
AM-CD_8
LM-ABS 84.867 16.134
LM-GT 77.227 24.310
LM-MKN 84.306 16.768
LM-WB 84.550 16.475
Continuous Read
Speech Recognizer
7. CONCLUSION AND FUTURE WORKS
Conclusion
 Amharic speech recognition application for dictation
 BRANA
 Main components : speech recognizer engine module and dictation
application module
 Sub modules of speech recognizer : Acoustic model module, language
model module and pronunciation dictionary module
 Front end user Interface
 90 min spontaneous speech corpus and 20 hours read speech corpus
 HMM approach
15
Concl…(Cont.)
Future Works…
 Automatic Error Correction
 Improve performance of speech recognizer
 Incorporate command and control
 Noise canceling
 Generic dictator
 Spontaneous speech corpus
16
Thank You!
Questions?
17

More Related Content

Similar to 8th Ethiopian ICT Conference Bazaar and Exhibition.pptx

Seminar
SeminarSeminar
4Developers 2015: Talking and listening to web pages - Aurelio De Rosa
4Developers 2015: Talking and listening to web pages - Aurelio De Rosa4Developers 2015: Talking and listening to web pages - Aurelio De Rosa
4Developers 2015: Talking and listening to web pages - Aurelio De Rosa
PROIDEA
 
Python | What is Python | History of Python | Python Tutorial
Python | What is Python | History of Python | Python TutorialPython | What is Python | History of Python | Python Tutorial
Python | What is Python | History of Python | Python Tutorial
QA TrainingHub
 
10 World’s Leading Speech or Voice Recognition Software That Can 3X Your Prod...
10 World’s Leading Speech or Voice Recognition Software That Can 3X Your Prod...10 World’s Leading Speech or Voice Recognition Software That Can 3X Your Prod...
10 World’s Leading Speech or Voice Recognition Software That Can 3X Your Prod...
nehachhh
 
Speech Recognition Technology
Speech Recognition TechnologySpeech Recognition Technology
Speech Recognition Technology
Aamir-sheriff
 
Concept of computer programming iv
Concept of computer programming ivConcept of computer programming iv
Concept of computer programming iv
Eyelean xilef
 
Computer Science Is The Study Of Principals And How The...
Computer Science Is The Study Of Principals And How The...Computer Science Is The Study Of Principals And How The...
Computer Science Is The Study Of Principals And How The...
Laura Martin
 
Generations Of Programming Languages
Generations Of Programming LanguagesGenerations Of Programming Languages
Generations Of Programming Languages
py7rjs
 
IRJET- Vocal Code
IRJET- Vocal CodeIRJET- Vocal Code
IRJET- Vocal Code
IRJET Journal
 
Computer Programming
Computer Programming Computer Programming
Computer Programming
Newreborn Incarnation
 
Computer
ComputerComputer
Instant speech translation 10BM60080 - VGSOM
Instant speech translation   10BM60080 - VGSOMInstant speech translation   10BM60080 - VGSOM
Instant speech translation 10BM60080 - VGSOM
sathiyaseelanm
 
B tech project_report
B tech project_reportB tech project_report
B tech project_report
abhiuaikey
 
IDE Code Compiler for the physically challenged (Deaf, Blind & Mute)
IDE Code Compiler for the physically challenged (Deaf, Blind & Mute)IDE Code Compiler for the physically challenged (Deaf, Blind & Mute)
IDE Code Compiler for the physically challenged (Deaf, Blind & Mute)
IRJET Journal
 
Hindi speech enabled windows application using microsoft
Hindi speech enabled windows application using microsoftHindi speech enabled windows application using microsoft
Hindi speech enabled windows application using microsoft
IAEME Publication
 
A Review On Speech Feature Techniques And Classification Techniques
A Review On Speech Feature Techniques And Classification TechniquesA Review On Speech Feature Techniques And Classification Techniques
A Review On Speech Feature Techniques And Classification Techniques
Nicole Heredia
 
An Application for Performing Real Time Speech Translation in Mobile Environment
An Application for Performing Real Time Speech Translation in Mobile EnvironmentAn Application for Performing Real Time Speech Translation in Mobile Environment
An Application for Performing Real Time Speech Translation in Mobile Environment
Association of Scientists, Developers and Faculties
 
Introduction to python
Introduction to pythonIntroduction to python
Introduction to python
Jaya Kumari
 
Designing the Workflow of a Language Interpretation Device Using Artificial I...
Designing the Workflow of a Language Interpretation Device Using Artificial I...Designing the Workflow of a Language Interpretation Device Using Artificial I...
Designing the Workflow of a Language Interpretation Device Using Artificial I...
IOSR Journals
 
Ayushi
AyushiAyushi

Similar to 8th Ethiopian ICT Conference Bazaar and Exhibition.pptx (20)

Seminar
SeminarSeminar
Seminar
 
4Developers 2015: Talking and listening to web pages - Aurelio De Rosa
4Developers 2015: Talking and listening to web pages - Aurelio De Rosa4Developers 2015: Talking and listening to web pages - Aurelio De Rosa
4Developers 2015: Talking and listening to web pages - Aurelio De Rosa
 
Python | What is Python | History of Python | Python Tutorial
Python | What is Python | History of Python | Python TutorialPython | What is Python | History of Python | Python Tutorial
Python | What is Python | History of Python | Python Tutorial
 
10 World’s Leading Speech or Voice Recognition Software That Can 3X Your Prod...
10 World’s Leading Speech or Voice Recognition Software That Can 3X Your Prod...10 World’s Leading Speech or Voice Recognition Software That Can 3X Your Prod...
10 World’s Leading Speech or Voice Recognition Software That Can 3X Your Prod...
 
Speech Recognition Technology
Speech Recognition TechnologySpeech Recognition Technology
Speech Recognition Technology
 
Concept of computer programming iv
Concept of computer programming ivConcept of computer programming iv
Concept of computer programming iv
 
Computer Science Is The Study Of Principals And How The...
Computer Science Is The Study Of Principals And How The...Computer Science Is The Study Of Principals And How The...
Computer Science Is The Study Of Principals And How The...
 
Generations Of Programming Languages
Generations Of Programming LanguagesGenerations Of Programming Languages
Generations Of Programming Languages
 
IRJET- Vocal Code
IRJET- Vocal CodeIRJET- Vocal Code
IRJET- Vocal Code
 
Computer Programming
Computer Programming Computer Programming
Computer Programming
 
Computer
ComputerComputer
Computer
 
Instant speech translation 10BM60080 - VGSOM
Instant speech translation   10BM60080 - VGSOMInstant speech translation   10BM60080 - VGSOM
Instant speech translation 10BM60080 - VGSOM
 
B tech project_report
B tech project_reportB tech project_report
B tech project_report
 
IDE Code Compiler for the physically challenged (Deaf, Blind & Mute)
IDE Code Compiler for the physically challenged (Deaf, Blind & Mute)IDE Code Compiler for the physically challenged (Deaf, Blind & Mute)
IDE Code Compiler for the physically challenged (Deaf, Blind & Mute)
 
Hindi speech enabled windows application using microsoft
Hindi speech enabled windows application using microsoftHindi speech enabled windows application using microsoft
Hindi speech enabled windows application using microsoft
 
A Review On Speech Feature Techniques And Classification Techniques
A Review On Speech Feature Techniques And Classification TechniquesA Review On Speech Feature Techniques And Classification Techniques
A Review On Speech Feature Techniques And Classification Techniques
 
An Application for Performing Real Time Speech Translation in Mobile Environment
An Application for Performing Real Time Speech Translation in Mobile EnvironmentAn Application for Performing Real Time Speech Translation in Mobile Environment
An Application for Performing Real Time Speech Translation in Mobile Environment
 
Introduction to python
Introduction to pythonIntroduction to python
Introduction to python
 
Designing the Workflow of a Language Interpretation Device Using Artificial I...
Designing the Workflow of a Language Interpretation Device Using Artificial I...Designing the Workflow of a Language Interpretation Device Using Artificial I...
Designing the Workflow of a Language Interpretation Device Using Artificial I...
 
Ayushi
AyushiAyushi
Ayushi
 

Recently uploaded

Microsoft - Power Platform_G.Aspiotis.pdf
Microsoft - Power Platform_G.Aspiotis.pdfMicrosoft - Power Platform_G.Aspiotis.pdf
Microsoft - Power Platform_G.Aspiotis.pdf
Uni Systems S.M.S.A.
 
How to use Firebase Data Connect For Flutter
How to use Firebase Data Connect For FlutterHow to use Firebase Data Connect For Flutter
How to use Firebase Data Connect For Flutter
Daiki Mogmet Ito
 
Choosing The Best AWS Service For Your Website + API.pptx
Choosing The Best AWS Service For Your Website + API.pptxChoosing The Best AWS Service For Your Website + API.pptx
Choosing The Best AWS Service For Your Website + API.pptx
Brandon Minnick, MBA
 
Cosa hanno in comune un mattoncino Lego e la backdoor XZ?
Cosa hanno in comune un mattoncino Lego e la backdoor XZ?Cosa hanno in comune un mattoncino Lego e la backdoor XZ?
Cosa hanno in comune un mattoncino Lego e la backdoor XZ?
Speck&Tech
 
Video Streaming: Then, Now, and in the Future
Video Streaming: Then, Now, and in the FutureVideo Streaming: Then, Now, and in the Future
Video Streaming: Then, Now, and in the Future
Alpen-Adria-Universität
 
Full-RAG: A modern architecture for hyper-personalization
Full-RAG: A modern architecture for hyper-personalizationFull-RAG: A modern architecture for hyper-personalization
Full-RAG: A modern architecture for hyper-personalization
Zilliz
 
Presentation of the OECD Artificial Intelligence Review of Germany
Presentation of the OECD Artificial Intelligence Review of GermanyPresentation of the OECD Artificial Intelligence Review of Germany
Presentation of the OECD Artificial Intelligence Review of Germany
innovationoecd
 
Driving Business Innovation: Latest Generative AI Advancements & Success Story
Driving Business Innovation: Latest Generative AI Advancements & Success StoryDriving Business Innovation: Latest Generative AI Advancements & Success Story
Driving Business Innovation: Latest Generative AI Advancements & Success Story
Safe Software
 
Essentials of Automations: The Art of Triggers and Actions in FME
Essentials of Automations: The Art of Triggers and Actions in FMEEssentials of Automations: The Art of Triggers and Actions in FME
Essentials of Automations: The Art of Triggers and Actions in FME
Safe Software
 
Let's Integrate MuleSoft RPA, COMPOSER, APM with AWS IDP along with Slack
Let's Integrate MuleSoft RPA, COMPOSER, APM with AWS IDP along with SlackLet's Integrate MuleSoft RPA, COMPOSER, APM with AWS IDP along with Slack
Let's Integrate MuleSoft RPA, COMPOSER, APM with AWS IDP along with Slack
shyamraj55
 
GenAI Pilot Implementation in the organizations
GenAI Pilot Implementation in the organizationsGenAI Pilot Implementation in the organizations
GenAI Pilot Implementation in the organizations
kumardaparthi1024
 
HCL Notes and Domino License Cost Reduction in the World of DLAU
HCL Notes and Domino License Cost Reduction in the World of DLAUHCL Notes and Domino License Cost Reduction in the World of DLAU
HCL Notes and Domino License Cost Reduction in the World of DLAU
panagenda
 
Building Production Ready Search Pipelines with Spark and Milvus
Building Production Ready Search Pipelines with Spark and MilvusBuilding Production Ready Search Pipelines with Spark and Milvus
Building Production Ready Search Pipelines with Spark and Milvus
Zilliz
 
Infrastructure Challenges in Scaling RAG with Custom AI models
Infrastructure Challenges in Scaling RAG with Custom AI modelsInfrastructure Challenges in Scaling RAG with Custom AI models
Infrastructure Challenges in Scaling RAG with Custom AI models
Zilliz
 
GraphRAG for Life Science to increase LLM accuracy
GraphRAG for Life Science to increase LLM accuracyGraphRAG for Life Science to increase LLM accuracy
GraphRAG for Life Science to increase LLM accuracy
Tomaz Bratanic
 
Mind map of terminologies used in context of Generative AI
Mind map of terminologies used in context of Generative AIMind map of terminologies used in context of Generative AI
Mind map of terminologies used in context of Generative AI
Kumud Singh
 
CAKE: Sharing Slices of Confidential Data on Blockchain
CAKE: Sharing Slices of Confidential Data on BlockchainCAKE: Sharing Slices of Confidential Data on Blockchain
CAKE: Sharing Slices of Confidential Data on Blockchain
Claudio Di Ciccio
 
Generating privacy-protected synthetic data using Secludy and Milvus
Generating privacy-protected synthetic data using Secludy and MilvusGenerating privacy-protected synthetic data using Secludy and Milvus
Generating privacy-protected synthetic data using Secludy and Milvus
Zilliz
 
National Security Agency - NSA mobile device best practices
National Security Agency - NSA mobile device best practicesNational Security Agency - NSA mobile device best practices
National Security Agency - NSA mobile device best practices
Quotidiano Piemontese
 
Your One-Stop Shop for Python Success: Top 10 US Python Development Providers
Your One-Stop Shop for Python Success: Top 10 US Python Development ProvidersYour One-Stop Shop for Python Success: Top 10 US Python Development Providers
Your One-Stop Shop for Python Success: Top 10 US Python Development Providers
akankshawande
 

Recently uploaded (20)

Microsoft - Power Platform_G.Aspiotis.pdf
Microsoft - Power Platform_G.Aspiotis.pdfMicrosoft - Power Platform_G.Aspiotis.pdf
Microsoft - Power Platform_G.Aspiotis.pdf
 
How to use Firebase Data Connect For Flutter
How to use Firebase Data Connect For FlutterHow to use Firebase Data Connect For Flutter
How to use Firebase Data Connect For Flutter
 
Choosing The Best AWS Service For Your Website + API.pptx
Choosing The Best AWS Service For Your Website + API.pptxChoosing The Best AWS Service For Your Website + API.pptx
Choosing The Best AWS Service For Your Website + API.pptx
 
Cosa hanno in comune un mattoncino Lego e la backdoor XZ?
Cosa hanno in comune un mattoncino Lego e la backdoor XZ?Cosa hanno in comune un mattoncino Lego e la backdoor XZ?
Cosa hanno in comune un mattoncino Lego e la backdoor XZ?
 
Video Streaming: Then, Now, and in the Future
Video Streaming: Then, Now, and in the FutureVideo Streaming: Then, Now, and in the Future
Video Streaming: Then, Now, and in the Future
 
Full-RAG: A modern architecture for hyper-personalization
Full-RAG: A modern architecture for hyper-personalizationFull-RAG: A modern architecture for hyper-personalization
Full-RAG: A modern architecture for hyper-personalization
 
Presentation of the OECD Artificial Intelligence Review of Germany
Presentation of the OECD Artificial Intelligence Review of GermanyPresentation of the OECD Artificial Intelligence Review of Germany
Presentation of the OECD Artificial Intelligence Review of Germany
 
Driving Business Innovation: Latest Generative AI Advancements & Success Story
Driving Business Innovation: Latest Generative AI Advancements & Success StoryDriving Business Innovation: Latest Generative AI Advancements & Success Story
Driving Business Innovation: Latest Generative AI Advancements & Success Story
 
Essentials of Automations: The Art of Triggers and Actions in FME
Essentials of Automations: The Art of Triggers and Actions in FMEEssentials of Automations: The Art of Triggers and Actions in FME
Essentials of Automations: The Art of Triggers and Actions in FME
 
Let's Integrate MuleSoft RPA, COMPOSER, APM with AWS IDP along with Slack
Let's Integrate MuleSoft RPA, COMPOSER, APM with AWS IDP along with SlackLet's Integrate MuleSoft RPA, COMPOSER, APM with AWS IDP along with Slack
Let's Integrate MuleSoft RPA, COMPOSER, APM with AWS IDP along with Slack
 
GenAI Pilot Implementation in the organizations
GenAI Pilot Implementation in the organizationsGenAI Pilot Implementation in the organizations
GenAI Pilot Implementation in the organizations
 
HCL Notes and Domino License Cost Reduction in the World of DLAU
HCL Notes and Domino License Cost Reduction in the World of DLAUHCL Notes and Domino License Cost Reduction in the World of DLAU
HCL Notes and Domino License Cost Reduction in the World of DLAU
 
Building Production Ready Search Pipelines with Spark and Milvus
Building Production Ready Search Pipelines with Spark and MilvusBuilding Production Ready Search Pipelines with Spark and Milvus
Building Production Ready Search Pipelines with Spark and Milvus
 
Infrastructure Challenges in Scaling RAG with Custom AI models
Infrastructure Challenges in Scaling RAG with Custom AI modelsInfrastructure Challenges in Scaling RAG with Custom AI models
Infrastructure Challenges in Scaling RAG with Custom AI models
 
GraphRAG for Life Science to increase LLM accuracy
GraphRAG for Life Science to increase LLM accuracyGraphRAG for Life Science to increase LLM accuracy
GraphRAG for Life Science to increase LLM accuracy
 
Mind map of terminologies used in context of Generative AI
Mind map of terminologies used in context of Generative AIMind map of terminologies used in context of Generative AI
Mind map of terminologies used in context of Generative AI
 
CAKE: Sharing Slices of Confidential Data on Blockchain
CAKE: Sharing Slices of Confidential Data on BlockchainCAKE: Sharing Slices of Confidential Data on Blockchain
CAKE: Sharing Slices of Confidential Data on Blockchain
 
Generating privacy-protected synthetic data using Secludy and Milvus
Generating privacy-protected synthetic data using Secludy and MilvusGenerating privacy-protected synthetic data using Secludy and Milvus
Generating privacy-protected synthetic data using Secludy and Milvus
 
National Security Agency - NSA mobile device best practices
National Security Agency - NSA mobile device best practicesNational Security Agency - NSA mobile device best practices
National Security Agency - NSA mobile device best practices
 
Your One-Stop Shop for Python Success: Top 10 US Python Development Providers
Your One-Stop Shop for Python Success: Top 10 US Python Development ProvidersYour One-Stop Shop for Python Success: Top 10 US Python Development Providers
Your One-Stop Shop for Python Success: Top 10 US Python Development Providers
 

8th Ethiopian ICT Conference Bazaar and Exhibition.pptx

  • 1. BRANA (ብራና): APPLICATION OF AMHARIC SPEECH RCOGNITION SYSTEM FOR DICTATION IN JUDICIAL DOMAIN Presenter: Bantegize Addis Adviser: Solomon Teferra June 05, 2015
  • 2. What I’m going to talk about… 1. Introduction… 2. Related Works…. 3. Speech database…. 4. Architecture… 5. Implementation… 6. Our Results…. 7. Conclusion & future works… 2
  • 3. 1. Introduction General Background  Automatic speech recognition gives us a new channel for communication with computers.  Speech technology is the technology of today and tomorrow.  It has practical implementations for both fun and serious works.  Mostly applied in command and control, data entry and retrieval, and dictation functions. 3
  • 4. Intro.(Cont.) Statement of the Problem  Dictation has been a common application area for ASRS for a long period.  Amharic is the second most-spoken Semitic language in the world .  It is the official working language of the FDRE.  To the best of our knowledge, there is no attempted works about application of speech recognition for dictation in Amharic. 4
  • 5. 2. Related works  Automatic speech recognition for Amharic 5 Author Type of recognizer Unit of recognition Recognition result Solomon Berhanu Isolated Consonant-Vowel (CV) Syllable Speaker dependent : 87.68% Speaker independent: 72.75% Speaker not involved in the training: 49.21% Kinfe Taddesse Isolated Phoneme Tri-phone Consonant-Vowel (CV) Syllable Speaker independent tri-phone (Test set I): 91.46 % Speaker independent tri-phone (Test set I): 77.87%
  • 6. Related…(Cont.) Author Type of Recognizer Unit of Recognition Recognition result Solomon Teferra Continuous Tri-phone Consonant-Vowel (CV) Syllable Speaker Independent tri- phone: 91.31% Speaker Independent CV: of 90.43% 6 Application of Amharic Speech Recognition Martha developed Amharic speech input interface to command and control Microsoft Word
  • 7. 3. Speech Corpus & annotation  Two types of Speech Corpus I. Spontaneous speech corpus II. Read speech corpus 7
  • 8. 4. The Architecture of BRANA 8
  • 9. 5. Implementation of BRANA Development Tools  JDK jdk1.7.0_05 with Eclipse  Sphinx-4  Sphinx Trainer  SRILM  cmuclmtk-0.7-win32 9
  • 10. Implementation…(Cont.) User Interface Implementation  The user interface of our system is responsible for:  editing rtf text documents  PPT(Push To Talk) functionality  displaying the uttered word or sentence hypothesis to the user  Developed using Java 10
  • 11. Implementation…(Cont.) 11 Push button to initiate recogntion GUI snapshot of BRANA(ብራና)
  • 12. Implementation…(Cont.) Python program implementation  Pronunciation Dictionary Python script program is implemented for generating a grapheme-based canonical pronunciation dictionary 12
  • 13. 6. Our results… 13 Recognition performance of spontaneous speech recognizer (In a Batch Recognition Mode) Acoustic Model Language Model Accuracy (%) WER (%) AM-CD_8 LM-ABS 49.537 55.093 LM-GT 49.769 56.019 LM-MKN 50.463 54.861 LM-WB 49.769 55.093 Spontaneous Speech Recognizer
  • 14. result…(Cont.) 14 Recognition performance of continuous read speech recognizer (In a Batch Recognition Mode) Acoustic Model Language Model Accuracy (%) WER (%) AM-CD_8 LM-ABS 84.867 16.134 LM-GT 77.227 24.310 LM-MKN 84.306 16.768 LM-WB 84.550 16.475 Continuous Read Speech Recognizer
  • 15. 7. CONCLUSION AND FUTURE WORKS Conclusion  Amharic speech recognition application for dictation  BRANA  Main components : speech recognizer engine module and dictation application module  Sub modules of speech recognizer : Acoustic model module, language model module and pronunciation dictionary module  Front end user Interface  90 min spontaneous speech corpus and 20 hours read speech corpus  HMM approach 15
  • 16. Concl…(Cont.) Future Works…  Automatic Error Correction  Improve performance of speech recognizer  Incorporate command and control  Noise canceling  Generic dictator  Spontaneous speech corpus 16