Real time Sign Language detection

Purpose:

Project primarily deals with TensorFlow object detection API,CNN,SSD mobilenet and OpenCV. This Project uses a Transfer learning based approach using pretrained models and tested using the datasets that we created for testing . Real time conversion done at the end where images are captured by camera.

Key Features:

• Preparing the train and test datasets

• Real time conversion

• Hand segmentation approaches

• Feature extraction approaches

• Gesture recognition approach

Design Considerations:

• CNN

• TensorFlow

• OpenCV

• CUDA(10.1)

• CuDNN(7.6) (For NIVIA GPU for faster training model)