A real time hand gesture detection and emoji translation using mediapipe, pillow and opencv. Currently, the following emojis are supported for both left and right hand.
- closed fist ✊
- I love you 🤟
- open palm ✋
- pointing up ☝️
- victory ✌️
I recommend you to create a new environment before installation. Current mediapipe only support python version 3.10.
conda create -n vidmoji python=3.10conda activate vidmojiconda install opencv pillow
pip install mediapipeCurrently, I don't support configuration over command line options. You can edit in the source code directly in vidmoji.py.
# Config
DETECTION_TIME = 0.3 # seconds to wait before showing emoji
NUM_HANDS = 2 # number of hands to detect
EMOJI_DURATION = 2.0 # seconds to show
EMOJI_SIZE = 50 # size of emoji
EMOJI_PATH = "./emojis/" # path to emojis folderTo run the application, just run the vidmoji.py and press q to quit.
python ./vidmoji.py`In online casual video meeting, users can raise hand, send emojis to use gestures instead of pressing buttons.
For the next step, I would like to implement emotion detection and emotions to emoji. This one is in progress. You can check in other branch.
Credits for Icon Pack