Hand_guesture_recognition

guessing the orintation of the hand and out putting the voice messages that are to be conveyed through the hardware using rasberry pi which is under the process.

##Requirements:

1. Anaconda - This installs python along with most popular python libraries including sklearn. If not already installed, install it from https://www.continuum.io/downloads .
2. Keras- Python library for Deep laerning.
3. open cv - Python library for playing with the images.

##Preparing dataset:

• Put all the unlabelled audio files in a folder named calls, or any other folder and update the name of folder in handguesture.py. -Then the preprocessed data will be created and stored in the data.npy file will be created along with the lables.npy -Then the model will be created and stored in the format of .h5 file for the furthur use.

##Output format:

• Out of all i have used 3 classes for the Convelutional model to determine which class the image depend upon.
• The out put format is always a softmax vector of probabilities and the maximum one too.
• We can apply the out put aany where as of now i am trying to apply it for understanding the signs of the dumb people.