Pakistan Sign Language Recognition

Project Title

Pakistan Sign Language Recognition

Project Area of Specialization Artificial IntelligenceProject Summary

The project will involve collecting a dataset, performing research and finally developing a web application in the light of that research.

The final deliverable which will be a web application is expected to recognize hand gestures from a video and output to the user its predictions. 

Project Objectives

• Research objectives
  • Study and evaluate existing methods for sign-language translation.
  • Developing a methodology for the solution of the problem.
  • Collecting datasets for training our algorithm.
  • Training the algorithm and drawing conclusions based on results.
• Academic Objectives
  • ???????Taking courses on Computer Vision.
  • Understanding the procedure for making such a web app and developing it.

Project Implementation Method

The discussed development methodology is rather general and not specific which is due to the fact that the dataset for the identified problem (Pakistan Sign Language gesture recognition) is neither available nor is there any universally accepted sign language. Therefore, a detailed methodology ( involving the names of the specific algorithms for image preprocessing as well as the detailed architecture of the CNN to be employed) is almost impossible without first collecting the required dataset and analyzing it to obtain methods that are most suitable to it. 

Following are the 3 major steps identified as for now:

1. Receiving a video from user

2. Uploading and processing the video on the server side (backend)

3. Displaying to the user the recognized gestures

Benefits of the Project

• Direct Customers/Beneficiaries of the project

  • Deaf-mute community

• Outputs Expected from the Project

  • A research-based technique for classification of signs in the Pakistan Sign Language

  • A web-based application which will take input from user in form of a video and provide the translation of the hand-gestures used in the video

Technical Details of Final Deliverable

The final deliverable consists of a software system made up of two software applications which are:

1. a user application
2. a processing application (which will process user input)

(1):

• will be a traditional client-side application of the client-server model.
• will accept a video input of hand gestures.
• will use a web API to send the video for processing.
• will receive the translation (of the hand gestures in the input video) through a web API.
• will present the recieved translation to its user

(2):

• will be a server-side application of the client-server model.
• will receive the video sent by the client-side.
• will apply the methodology to translate the hand gestures in the video.
• will send translation to (1)

Final Deliverable of the Project Software SystemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Industry, Innovation and Infrastructure, Reduced InequalityRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	74302
NVIDIA GEFORCE RTX 3060 TI	Equipment	1	64302	64302
Installation price for GPU	Miscellaneous	1	10000	10000