CNN model trained to be able to classify gestures in order to build a more inclusive gaming experience for everyone.
The model can be tested, independently or with the PyAutoGUI output through the test-model.ipynb file.
Libraries needed: numpy, cv2, tensorflow, pyautogui, opendatasets, pandas, sklearn, os.
ML_GestPlay.ipynb- Primary notebook for the model training. Model was trained on Google Colab to take advantage of their T4 GPU runntime.
test-model.ipynb- Primary testing notebook. Contains all the code for evaluating the model, as well as evaluating how feasible it is to integrate model output to keyboard inputs.
gesture_mode.h5- The model itself.
./04/- Directory with images used during the testing process. These are from the dataset that we split for test and train.