Translation Specula

Unique Registration Number: GSInC-I-000321

Innovator Name:

1. Shivam Raikar
2. Jesven Gomes
3. Vinayak Naik
4. Vinayak Halshikar

Name of a Mentor:-

1. Prof. Vaibhavi Naik
2. Prof. Manjusha Sanke

Name of School/College/Startup/Organisation: Shree Rayeshwar Institute of Engineering and Information Technology

Contact No:

• +91 9923679093
• +91 9890658995

Contact Email:

• shivam.mac100@gmail.com
• jesvengomes23@gmail.com

Project Objective:

• The Project Objective is to design and develop a real time language translation glasses that enables the spectacle wearers to have instant and accurate translation of the spoken language that is displayed before their eyes.

Abstract:

• Language barriers cause problems in our global world. People can’t understand each other in education, business, travel, and everyday life. Translating languages can be hard, and conversations require translators or devices, making eye contact difficult. This project created real-time translation glasses(spectacles) to break down these barriers. The glasses use advanced technology to translate spoken words instantly. They show the translation as text or play it in your preferred language. These glasses could be very helpful in many situations, allowing people to understand each other during business meetings or casual conversations. The project shows that these glasses can work and have a big impact. They are easy to use and combine technology well. This could make communication better and help people around the world understand and work together.

Project Outcome/result/findings:

• The Project was implemented and the spoken language is translated and displayed on the screen of the glasses.
• Findings:
  1. Due to the focal length of the eye sight we have made the display adjustable.
  2. We first took the Spanish language dataset that contained 3000 rows of reference(english) and translated (Spanish) data. We used the BLEU score to determine the accuracy of the model.
     • Calculated Accuracy of the model before training using nltk library to compute BLEU score. The accuracy score of the model is shown in the below table:

Innovative Approach:

• Real time Translation of the language that is spoken is displayed on the glasses.