AI Real-Time Sign Language Gesture Recognition System
Real-time hand gesture and sign language recognition system converting American/Pakistani Sign Language into text and synthetic speech using MediaPipe and LSTM.
### AI Real-Time Sign Language Translation System (Audio/Text Output)
An assistive AI engineering project that bridges the communication gap for deaf and mute individuals. Using MediaPipe to extract 21 3D hand landmarks in real time and a Temporal Long Short-Term Memory (LSTM) Recurrent Neural Network, the system translates dynamic sign gestures into English/Urdu text and natural spoken voice output.
#### 🧠 Architecture & Tech Stack:
- **Landmark Extraction**: Google MediaPipe Hand Holistic Tracker (63 spatial coordinates per frame)
- **Sequence Classifier**: Multi-layer Bidirectional LSTM in TensorFlow / Keras
- **Voice Synthesis**: pyttsx3 / gTTS Text-to-Speech Engine
- **Interface**: Clean, low-latency Tkinter GUI / Streamlit web dashboard
#### 📦 Complete Deliverables:
- Full Python Source Code with real-time video inference loop
- Custom dataset collector script & pre-trained LSTM gesture model
- Complete Final Year Project Documentation & Defense Slides
What's Included in the Package:
- Full Python Source Code & Inference Scripts
- Pre-trained LSTM Sign Language Model Weights
- Custom Gesture Recording & Training Utility
- Complete IEEE Project Report (Word DOC & PDF)
- Defense Presentation Slides (PPTX)
- Instant Technical Support
Key Features
Specifications
Community Questions & Answers
Have a question about compatibility, code, or delivery? Ask our engineering team or see what other Pakistani makers asked.
No questions asked yet for this item
Be the first student or maker to ask a question!