Currency Detection Using Android for Blind

Project Title

Currency Detection Using Android for Blind

Project Area of Specialization Artificial IntelligenceProject Summary

People having a visual impairment can have trouble/difficulty in several tasks like detection of currency notes. As part of this project the student will develop an Android app that will help a visually impaired person to recognize and detect the denomination of a specific currency note. The concerned person will give a command to the app by speaking to it to open the phone camera of a smartphone which then take a picture of the currency note. The user will then be informed of the denomination of the currency note by voice. This project will use speech to text conversion for transferring the command of the visually impaired person to the android device. The android app will use text to speech conversion to detect the denomination of the note and then will convert the text value to speech to the user for hearing. The application will use OpenCV using Machine learning for classification and detection of currency based on images or paper using smartphone camera.

Project Objectives

To solve our proposed problem, we need to develop an android application as a solution for implementation of currency recognition model and provide an easy to use interface and highly responsive design. To make it highly responsive and easy to use, we need to integrate a voice recognition and response module. Also, we need to update the dataset of the model frequently to improve on accuracy.

Project Implementation Method

Technologies that will be used in this project are Python which is an easy to use platform for application development for Android. OpenCV for image recognition and GitHub as a code repository for modules. For UI design, Adobe XD will be used. Voice recognition and text to speech modules will also be added to offer ease of use for the blind person. The picture of currency note will be automatically accepted for recognition only if the whole note is inside the picture frame. This will also be programmed into the application. We will also train our recognition model on a high spec GPU-enabled machine.

Benefits of the Project

This application will benefits blind people in their day to day life to recognize the currency notes.

Provide data for future research on detecting fake currency notes for blind as well as banking purposes.

Technical Details of Final Deliverable

Android Application, Currency Recognition Model and Dataset.

Technologies that will be used in this project are Python which is an easy to use platform for application development for Android. OpenCV for image recognition and GitHub as a code repository for modules. For UI design, Adobe XD will be used. Voice recognition and text to speech modules will also be added to offer ease of use for the blind person. The picture of currency note will be automatically accepted for recognition only if the whole note is inside the picture frame. This will also be programmed into the application. We will also train our recognition model on a high spec GPU-enabled machine.

Final Deliverable of the Project Software SystemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	72000
Nvidia Geforce GTX 1650 4GB	Equipment	1	69000	69000
HD Camera for Dataset Pictures	Miscellaneous	1	3000	3000