Scribd
Upload
3 views

AI Sign Language Detector Updated

The document presents an AI-powered real-time sign language detector focusing on three ASL signs using Python, OpenCV, and MediaPipe. It addresses the challenges of existing systems by providing a lightweight solution for the deaf and hard-of-hearing community, utilizing hand landmark detection for improved accuracy. Future enhancements aim to expand the recognition capabilities and deploy the system as a mobile or web application.

Uploaded by

xogix31164
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
3 views

AI Sign Language Detector Updated

The document presents an AI-powered real-time sign language detector focusing on three ASL signs using Python, OpenCV, and MediaPipe. It addresses the challenges of existing systems by providing a lightweight solution for the deaf and hard-of-hearing community, utilizing hand landmark detection for improved accuracy. Future enhancements aim to expand the recognition capabilities and deploy the system as a mobile or web application.

Uploaded by

xogix31164
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 12

AI-Powered Sign Language

Detector
Hackathon Presentation
Presented by: Sham
Date: 2025
Introduction
• • A real-time sign language detector using
Python, OpenCV, and MediaPipe.
• • Focuses on recognizing three ASL signs: A, B,
and L.
• • Uses hand landmark detection instead of full
image processing.
• • Bridges the communication gap for the deaf
and hard-of-hearing community.
Problem Statement
• • Sign language communication is essential for
the deaf and hard-of-hearing community.
• • Many existing recognition systems require
large datasets and high computational power.
• • Real-time detection with high accuracy
remains a challenge.
• • A lightweight and efficient solution is
needed for better accessibility.
Technology Stack
• • Programming Languages: Python, Java
• • Libraries and Frameworks:
• - OpenCV (image processing)
• - MediaPipe (hand landmark detection)
• - Scikit-Learn (machine learning model
training)
• - Java (for Android integration and
application development)
Approach & Methodology
• 1. Data Collection: Capturing hand gesture
images.
• 2. Preprocessing: Extracting hand landmarks.
• 3. Model Training: Using a Random Forest
classifier.
• 4. Testing & Evaluation: Measuring accuracy.
• 5. Real-Time Testing: Implementing with
webcam.
Data Preparation
• • Collect diverse hand gesture samples to
ensure accuracy.
• • Use different angles, hand sizes, and lighting
conditions.
• • Preprocess images to extract hand
landmarks using MediaPipe.
Model Selection & Training
• • Why Random Forest?
• - Simple yet effective for gesture
classification.
• - Works well with extracted hand landmark
features.
• • Training Process:
• - Features: Hand landmark coordinates.
• - Train-Test Split: Evaluating generalization
performance.
Real-Time Implementation
• • Integrated with webcam to detect signs in
real-time.
• • Uses MediaPipe’s efficient hand tracking.
• • Provides instant feedback on detected signs.
Key Insights & Learnings
• • Data diversity improves model accuracy.
• • Landmark detection simplifies classification
and reduces computational cost.
• • Iterative testing helps refine performance.
• • Real-time sign recognition bridges theory
with practical application.
Challenges & Future Scope
• Challenges:
• • Adapting to different hand orientations and
backgrounds.
• • Improving accuracy in real-time conditions.

• Future Enhancements:
• • Expanding to recognize full ASL alphabet.
• • Deploying as a mobile or web application.
Conclusion
• • Successfully developed a real-time ASL sign
detector.
• • Demonstrates the potential of AI in
accessibility solutions.
• • Future improvements will enhance usability
and scalability.
Thank You!
• • Questions? Reach out to us!
• • Contact: [Your Email/LinkedIn]
• • Let's build a more inclusive future together!

You might also like