Speech Rectification System

According to a survey conducted in Pakistan, 14 percent of Pakistanis suffer from speech disorders. Organizations like Pakistan Speech and Language Pathologist Association (PSLPA) are working to help children and adults with Speech Sound Disorder (SSD). Going through several training programs, peopl

Project Title

Speech Rectification System

Project Area of Specialization

Artificial Intelligence

Project Summary

According to a survey conducted in Pakistan, 14 percent of Pakistanis suffer from speech disorders. Organizations like Pakistan Speech and Language Pathologist Association (PSLPA) are working to help children and adults with Speech Sound Disorder (SSD). Going through several training programs, people can overcome this problem. There is a need for a support system to deliver better articulation instruction and practice in special education classes for children with an articulation disorder. Although many systems are available still there is a need for a system that provides real-time assistance to understand their speech and communicate clearly.

Speech articulation-disorders are characterized by mispronouncing phonetic units by deleting, distorting, substituting, or inserting sounds that make speech difficult to understand. So, we will develop a system that converts the disordered speech into an ordered one by the development of acoustic models and discovering articulation with the help of Data modelling, Artificial Intelligence, and Deep learning algorithms.

This application is designed to enable those who suffer from speech disorders to train their communication capabilities easily and entertainingly. Neurolinguistics will be used to understand how speech is organized in the brain before its delivered. We will be using Natural Language Processing algorithms also to understand the Urdu Language according to context, grammar, and vocabulary. It is based on real-time data receiving, processing, and correction.

Project Objectives

We will develop an application for the users that will be easily available on the Android play store for the potential users.

Objectives of our system are:

• To develop an easy to access application for the people with Speech Articulation Disorder, that will be helpful for children and adults alike who are facing difficulty in communication due to articulation disorders.
• Users will be able to speak in the Urdu language and the application will effectively recognize it.
• Determine the incorrect words, reconstruct the sentence with corrected words.
• Convert the regenerated speech into audio in real-time.

The system will keep the record of the person to see the improvement patterns in speech like how many articulation or phonology mistakes he/she made.

Project Implementation Method

This app will recognize Urdu Language with the help of Google APIs. Once the language is detected then we convert it into text and further into sentences and then perform correction of faulty words by using Language Model, Probabilistic and Natural Language Model. After that we will pass this input through LSTM network which is a type of Artificial Neural Networks then we perform k-fold Validation technique with text and get the correct sentences. After that we will receive output in the form of rectify speech then we trained that resulted speech by using Precision, Recall and F-Measures techniques and in the last we will compare the correct speech with realistic speech. The app will also keep record of the person’s speech corrections by using Artificial intelligence algorithm so we can track the progress of the user.

Benefits of the Project

• This system will help people to communicate easily because it transfers their disordered speech into correct speech.
• Users will be able to speak in the Urdu language which will be the greatest advantage.
• With the help of this app, users can easily control devices and create documents by speaking because it allows documents to be created faster because the software generally produces words as quickly as they uttered, which is usually much faster than a person can type.

It will reduce the lack of confidence in people with Speech Sound Disorder (SSD).

Technical Details of Final Deliverable

• Implementation will be done in python language.
• Detecting the Urdu language with API and interpret the context with the help of AI algorithms.
• Machine Learning algorithm will be used.
• A new deep learning algorithm will be developed to understand the speech in context.
• Using deep learning, the system will comprehend, compare, and correct the speech recognized by API.

Filter the correct sentence using AI and output the corrected sentence by the same API in speech form.

Final Deliverable of the Project

Software System

Core Industry

Medical

Other Industries

IT , Health

Core Technology

Artificial Intelligence(AI)

Other Technologies

Others

Sustainable Development Goals

Good Health and Well-Being for People

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com