BAYES WATCH

TECHNOLOGY

STUDIO SHODWE

HOME

SERVICE

ABOUT US

CONTACT US

Theme : Community Upliftment

Page 01

WHY VIDYA NETRA?

STUDIO SHODWE

HOME

SERVICE

ABOUT US

CONTACT US

India has over 9 million visually impaired individuals, including more than 1 million children. Despite hosting 20% of worlds visually impaired population.

​

Access to inclusive education is a major challenge:

• Most blind students rely on human scribes.
• India has a Braille Literacy of mere 1%.
• Textbooks in braille or audio format are scarce.
• Limited exposure to modern educational tools.

​

Blind students deserve independence, dignity, and equal access to learning.

Page 02

OUR PROPOSED SOLUTION

HOME

SERVICE

ABOUT US

CONTACT US

• PDF/Text to Audio & Braille Converter:
• Converts digital text (PDFs, etc.) into synthesized audio and digital Braille formats (e.g., BRF files).
• Makes existing written materials readily accessible for auditory learners and Braille reader users.
• Novel Braille Learning via Spatial Audio:
• Teaches Braille character patterns using distinct spatial audio cues corresponding to dot positions.
• Offers an intuitive, non-visual method to supplement traditional Braille learning techniques.
• AI-Powered Accessible Exam Portal:
• Features secure face/voice login, presents questions auditorily, and accepts dictated answers via voice.
• Enables independent, scribe-free examination for students with visual or impairments.
• Video Content Interaction (Transcript + Audio Q&A using RAG):
• Transcribes video content and uses Retrieval-Augmented Generation (RAG) for interactive Q&A.
• Allows users to ask questions about video content (via voice/text) and receive targeted audio/text answers derived from the transcript.

​

Page 04

HOME

SERVICE

ABOUT US

CONTACT US

• Processing Pipeline: Automatically identifies input (Image, PDF, Audio), uses OCR/STT for text extraction, analyzes visuals with AI Vision, and converts all content into accessible formats including high-quality TTS audio and Braille embeddings.

​

• Q&A Assistant (RAG): Allows students to ask natural language questions about their uploaded course materials, receiving contextually relevant answers generated by AI using Retrieval-Augmented Generation based only on that specific content.

​

• Secure Exam Accessibility: Provides secure, AI-powered conversion of question papers into accessible formats (TTS, Braille-ready, visual descriptions) and enables real-time AI dictation (STT) for students to answer within a secure testing environment

​

• Biometric Authentication: Enables secure and convenient login/verification using integrated AI-powered facial recognition and voice authentication.

IMPLEMENTATION DETAILS + TECH STACK

TECHSTACK

​

1. huggING FACE

2. STREAMLIT

3. GTTS

4. OLLAMA

5. QWEN1.5-7B

6. LANGCHAIN

7. CLOUDINARY

8. DOCLING

9. PILLOW

FUTURE WORK

STUDIO SHODWE

HOME

SERVICE

ABOUT US

CONTACT US

• Implement a voice-driven AI agent (using Natural Language Understanding) that allows users to navigate the entire VidhyaNetra platform and control its features using natural language commands, rather than remembering specific keywords.

​

​

Page 06

NAVIGATION

• For both converting diagrams and images into textual formats and then extracting key representable details from them, this process enables a deeper semantic understanding of visual content. These extracted insights can then be transformed into immersive binaural audio representations, making complex diagrams more accessible and engaging—especially for auditory learners or visually impaired users.

​

ADVANCED BINAURAL AUDIO

• Translate simple charts and graphs (e.g., bar charts, line graphs) into an interactive auditory experience. Users could "scan" across the data using voice commands or simple controls, hearing changes in pitch, volume, or spatial location of sound to represent data values and trends

DATA ANALYTICS

THANK YOU

STUDIO SHODWE

HOME

SERVICE

ABOUT US

CONTACT US

Page 10