The Real-Time Sign Language Alphabet Recognition System stands as a remarkable achievement in bridging communication gaps for hearing-impaired individuals and fostering societal inclusivity. Developed as a research project at Doğuş University, under the Software Engineering department, this groundbreaking system highlights the transformative power of machine learning and computer vision in addressing real-world challenges. Guided by principles of innovation and societal benefit, the project showcases how technology can touch lives and build a more inclusive world.
With a recognition accuracy exceeding 90%, the system seamlessly identifies sign language alphabet gestures and translates them into corresponding letters in real time. By utilizing cutting-edge techniques like deep learning, this platform ensures highly reliable and efficient recognition—even for dynamic gestures like “J” and “Z.” It combines technical brilliance with practicality, serving as both an educational aid for sign language learners and a valuable communication tool for hearing-impaired individuals.
More than just a piece of software, this project embodies the spirit of collaboration and societal awareness, uniting researchers, academics, and technology enthusiasts in a shared goal: enabling independence for hearing-impaired individuals and fostering greater empathy and understanding within communities. Its user-friendly design makes it accessible to individuals of all ages and technical backgrounds, proving that inclusivity and simplicity can go hand in hand.
- Real-Time Recognition: Instant identification of sign language gestures via camera input.
- High Accuracy: Recognizes letters with an accuracy rate exceeding 90%, ensuring reliable results.
- Dynamic Gesture Handling: Supports complex and dynamic gestures for letters like "J" and "Z".
- Optimized for Accessibility: Simple and intuitive interface designed for users of all ages and technical backgrounds.
- Cross-Societal Impact: Helps raise awareness and understanding of sign language while fostering inclusivity.
- Empower hearing-impaired individuals by simplifying communication in their daily lives.
- Provide a reliable educational tool for learning sign language.
- Promote inclusivity by increasing societal awareness of sign language.
- Framework: TensorFlow (Deep Learning Framework).
- Programming Language: Python.
- Tools and Libraries: OpenCV for computer vision tasks and Numpy for numerical operations.
- Model Architecture: Convolutional Neural Network (CNN) designed for efficient gesture recognition.
- Dataset: Custom-built, diverse dataset including static and dynamic gestures to ensure robust performance.
- Input: The system uses the camera to capture hand gestures performed by the user.
- Preprocessing: Captured images are preprocessed for consistent recognition (e.g., resizing, normalization).
- Model Prediction: The machine learning model processes the input and matches it to corresponding letters.
- Output: Recognized letters are displayed on the screen in real-time, optionally with voice feedback.
-
Clone the Repository: git clone https://github.com/YourUsername/SignLanguagePython.git
-
Navigate to the Project Directory: cd RealTimeSignLanguageRecognition
-
Install Dependencies: pip install -r requirements.txt
-
Run the Application: python main.py
- Accuracy: Achieved a recognition rate exceeding 90% on both training and validation datasets.
- Dynamic Gesture Recognition: Successfully implemented dynamic gesture recognition for letters such as "J" and "Z", traditionally considered challenging.
- Optimized for Low-End Devices: Runs efficiently on low-resource machines, making it accessible to a broader audience.
- User Feedback: Extensive user testing highlighted the system's simplicity, accuracy, and practical utility.
- Ayşe Ceren Doğan
- Cemre Dağ
- Emir Ekrem Kaya
- Hatice Uçar
This project is licensed under the MIT License. See the LICENSE file for details.