Arduino Sign Language (ASL)

Ever wanted to be able to convert someone signing in ASL to

characters. With Arduino Sign Language, you can!

This project allows for people who do not understand American Sign Language themselves

to read what someone who is signing is trying to say. The project consists of a glove with

multiple sensors on it to detect the different positions of the hand along with the movement

that it is doing. Flex resistors were attached to each finger and could sense how curled

the finger is. An IMU (gyroscope/accelerometer) was connected to the top of the palm

to detect movement. The data was collected for each letter into a CSV file then fed into

a TensorFlow neural net as training. The model was exported after training then it was

used to predict the new data from the sensors when actually using the project. The model

was able to achieve around 90% accuracy in validation data.