Hand Pose Estimation and Hand recognition with Motion Prediction
Author: YiMing Chen
Author: Hengchen Zhou
Author: Mingyu Jiang
-
hand pose estimation with prediction.ipynb: This ipynb file is used for data importing, gesture prediction, data validation and evaluation.
-
hand pose estimation.ipynb: This ipynb file is used for data importing, model training and hand pose estimation
-
hand gesture recognition.ipynb: This ipynb file is used for recognize the gestures in images
-
hand gesture recognition2.ipynb: This ipynb file combine the pose estimation model with the gesture recgnition model to judge the current gesture in image.
-
data.py: This script is used for data processing part, it can pick up the hands part out of the image and compress the images resolution from 1080P to 64 * 64 and 32 * 32.
-
gesture_label.py: This is a python script we use to label the picture with gesture status and stored into Json format.
-
gesture_data.py: This is a python script we use to combine the processed image file with the gesture label json file to produce the data file for gesture recognition training and testing.
All the working scrips are in the Jupyter folder, please go straight to the Jupyter folder to execute *.ipynb.
You need to download below all the *.save data files and put them in the same directory the *.ipynb files in (/Jupyter).
- gesture-32data.save: https://drive.google.com/file/d/1klVnOfdhcI7j8raNxoBS6VehWIu_KsCp/view?usp=sharing
- test-32data.save: https://drive.google.com/open?id=1KOGMOF8CQX6EjPoj6HMT1kx8Gh6um-6S
- test-32data-fastward-onlylabel.save: https://drive.google.com/open?id=1gVuiHKFY_PkyVgGgOKvWD3XwNzYvDHYh
We develop and debug our project on the Google Colab environment.
All the file is written in python3