🤟 LESCOT

LESCO Sign Language Recognition Glove

LESCOT is an assistive technology initiative developed by students from Colegio Técnico Profesional Don Bosco with the mission of reducing communication barriers between deaf and hearing people in Costa Rica.

The project focuses on the development of an intelligent glove capable of recognizing Costa Rican Sign Language (LESCO) through wearable sensors and machine learning, translating hand gestures into understandable digital output.

🎯 Project Mission

LESCOT aims to promote inclusion, accessibility, and equal rights by providing a technological solution that facilitates communication with the deaf community and supports the principles of Costa Rican Law No. 9822, which promotes equality and inclusion for people with disabilities.

“Inclusion means accepting the diversity around us.”

👨‍💻 Role & Contributions

Lead Developer – Embedded Systems & Machine Learning

• Designed and prototyped the LESCOT wearable glove
• Integrated accelerometers and sensor multiplexing systems
• Collected and labeled LESCO gesture datasets
• Developed Python-based machine learning models for gesture recognition
• Implemented real-time gesture classification and translation
• Participated in testing and validation with LESCO users
• Contributed to public demonstrations at EXPOTEC 2023 & 2024

🧠 System Architecture

Hardware Components

• Accelerometers: MSA311 (3-axis, I2C)
• Multiplexer: PCA9548A (I2C device management)
• Microcontroller: QT Py ESP32-S3 / ESP32-WROOM-32E
• Wireless Communication: Bluetooth & Wi-Fi
• Power: Rechargeable portable battery
• Design: 3D-modeled glove module (Fusion 360)

Processing Units (Flexible Architecture)

• Raspberry Pi
• Jetson Orin Nano
• Laptop / external processing unit

This modular approach allows LESCOT to adapt to different environments and deployment scenarios.

🧪 Software & Machine Learning Pipeline

1. Data Collection

   • Accelerometer data captured from hand movements
   • Multiple repetitions per gesture
   • Data collected from different users

2. Signal Processing

   • Noise filtering
   • Normalization and calibration
   • Feature extraction (motion patterns)

3. Model Training

   • Algorithms: SVM and Neural Networks
   • Tools: Python, Scikit-learn, TensorFlow
   • Train/Test split with cross-validation
   • Achieved 85%+ accuracy on trained gestures

4. Real-Time Recognition

   • Low-latency inference
   • Confidence scoring
   • Continuous gesture detection

5. Output

   • Text-based translation
   • Text-to-speech support

✨ Key Features

• 🤟 Real-time LESCO gesture recognition
• 📡 Wireless operation (Bluetooth / Wi-Fi)
• 🔁 Expandable gesture vocabulary
• 🧤 Portable and wearable design
• 🖥️ User-friendly digital interface
• 🧩 Modular and scalable architecture

🧠 Research & Background

LESCOT was developed after analyzing existing sign language translators, including:

• PIELS Translator (Costa Rica – TEC)
• UCLA Sign Language Gloves (ASL)
• Lenguantec (Tecnológico de Monterrey)
• SignAloud (Lemelson-MIT)

This research helped define a locally adapted solution focused on LESCO, addressing the lack of Costa Rica–specific assistive technologies.

📈 Impact & Recognition

• Demonstrated at EXPOTEC 2023 and EXPOTEC 2024
• Recognized by Emerson as Best Industrial Electronics Project (10th grade)
• Positive feedback from users and educators
• Raised awareness about LESCO and accessibility technologies
• Promoted STEAM education with social impact

🚀 Future Enhancements

• Expand vocabulary to full phrases and sentences
• Integrate computer vision for hand pose estimation
• Develop a mobile application
• Improve miniaturization and battery efficiency
• Collaborate with LESCO educators and institutions
• Deploy LESCOT modules in public service locations

🧩 Skills Demonstrated

Technical Skills

• Embedded systems design (ESP32)
• Sensor integration and I2C communication
• Machine learning model development
• Python-based data processing
• Wearable device prototyping

Human-Centered Skills

• Accessibility-focused design
• Empathy-driven engineering
• Collaboration with end-user communities
• Technical documentation and public presentation

LESCOT demonstrates how technology, when designed with empathy and purpose, can become a powerful tool for inclusion, accessibility, and social transformation.