Urdu Speech to Sign Converter

Human Interaction is the sole of a society. With deaf as its part but limited in their interaction with others they feel isolated. This project focuses on this problem through a recognition process of predefined Urdu sentences that are sampled under varying conditions which will enable a normal pers

Project Title

Urdu Speech to Sign Converter

Project Area of Specialization

Artificial Intelligence

Project Summary

Human Interaction is the sole of a society. With deaf as its part but limited in their interaction with others they feel isolated. This project focuses on this problem through a recognition process of predefined Urdu sentences that are sampled under varying conditions which will enable a normal person to communicate with a deaf person. The database thus formed involves the basic daily sentences essential for communication is used to train the network up to a maximum efficiency. This network then on the basis of recognition algorithms involved trains itself which is then utilized to recognize the input from a speaker by giving out the most suitable output.

Project Objectives

1. Recognition of Pre-defined Urdu Sentences.
2. Conversion of Recognized Sentences into Relevant Signs.
3. Display of Signs on a Screen.

Project Implementation Method

We first start off with the uploading of the speech sentences data set following with the labels of the specified sentences to be recognized. Then we split the data in to training, validation and testing sets. After this we add the background noise files to make the system even more reliable. Then data augmentation is applied on the system which helps to increase the data set sffective size and help prevent the system from over-fitting. We then apply the convolutional neural network that involves an over all of 24 layers with filters in between. Now the training of the system starts which depending on the system specifcations takes almost an hour. After this the network thus formed is  available for evaluation. Further it is then integrated with the sign language videos which will be displayed as the output in response to the recognized input.

Benefits of the Project

One of the many uses involve:

1. Deaf Schools Teaching.
2. Household Interaction.
3. Offices
4. Conference
5. Hospitals

Technical Details of Final Deliverable

The system overall involves softwares like the voice recorder , MATLAB 2018b and video camera for the videos recording

Final Deliverable of the Project

Software System

Type of Industry

Education

Technologies

Artificial Intelligence(AI), Others

Sustainable Development Goals

Good Health and Well-Being for People, Quality Education, Reduced Inequality

Required Resources

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