1. Vineet Kumar Kankerwal -
Role: Created Data for training.
2. Tanisha Kriplani -
Role: Trained Model
3. Yashasvi -
Role: Prediction of Data
4. Priyam Gupta -
Role: Collect samples of data and documentation.
1.Neural Network Framework: TensorFlow Keras API
2.Optimization Algorithm: Adam (Adaptive Moment Estimation)
3.Image Processing Library: OpenCV (CV2)
4.Hand Detection Library: HandTrackingModule (Python)
The project focuses on developing a sign language translator that detects Indian Sign Language alphabets and translates them into English alphabets. The system employs a Convolutional Neural Network (CNN) with Adam optimization for efficient learning. Utilizing the TensorFlow Keras API, the model is trained for image classification, with real-time hand detection achieved through OpenCV and the HandTrackingModule Python library. The core idea is to bridge communication gaps between differently-abled individuals and others, while also promoting awareness of Indian Sign Language.
1. Dependencies Installation:
Ensure you have Python installed on your system.
Install required libraries using:
pip install tensorflow opencv-Python
2. Clone the Repository:
Clone this machine to your local machine
git clone [repository_url]
3. Navigate to Project Directory:
Change Directory to project folder
cd [Project_Folder]
4. Run the Application:
Execute the main script to run the sign language translator:
python prediction.py
Our core code captures video from the webcam, detects hands using the HandTrackingModule, and extracts the hand region. It resizes the hand region to a fixed size, feeds it into a sign language classifier ("trainedModel.h5"), and displays the predicted letter on the video feed. The program continuously runs in a loop, updating the predictions as the hands gesture different letters.
we are unable to upload data samples because there is too huge data that it is giving error while uploading. data more than 2.5GB