Computer Vision - Rock, Paper, Scissors

Work in Progress

Rock, paper, scissors using Teachable Machine to collect images to create a model that recognises hand gestures.

Creating the Model

• First step was to create the image model using Teachable Machine. Using the built in webcam, recording specific hand gestures directly to the website, between 300 - 450 images, infront of a plain backgroud and without my face in the same frame.
• Creating four different classes for the three specific hand gestures associated with the game and allowing those images to train the machine. This took a couple of minutes.
• Once the machine has been trained, the model is ready to be previewed and downloaded.

Creating the game

• Once the model has been downloaded to a .h5 file, the game play can be written.
• Using python to create a simple rock, paper, scissors game where the computers choice is a random guess and you are able to input your guess before the computers guess is revealed.

def play():
    player = get_user_choice()
    computer = get_computer_choice()
    print(get_winner(computer, player))

• After testing the game, it was clear that the code was not recognising the user/player input when the first letter of the word was not capitalised.
• A simple adjustment of using the .capitalize() method to ensure that what ever the input from the use is, the first letter will be capitalised.