README.md

Sign Language to speech using mediapipe and RNN with keras

Frameworks & liberaries used

- MediaPipe by Google.

- TensorFlow

- Google Text To speech

Dataset collection process

project demonstration 👀

Turn volume up 🔊🔊

Untitled.mp4

Packages Required

python3 -m pip install tensorflow opencv-python mediapipe sklearn matplotlib

For Text to speech

python3 -m pip install gtts

To play resultant mp3 file from gtts

mpg123 is used in a linux machine

sudo apt update
sudo apt install mpg123

text to speech function implementation in linux 🐧

processes = []

def text_to_speech(sentence):
    output = gTTS(sentence, lang='en', slow=False)
    output.save("output.mp3")
    process = subprocess.Popen("mpg123 output.mp3", shell=True)
    processes.append(process)

output = [p.wait() for p in processes]

text to speech function implementation in windows 🪟

NOTE ⚠️: in windows you don't need mpg123 to run .mp3 files in a windows machine

processes = []

def text_to_speech(sentence):
    output = gTTS(sentence, lang='en', slow=False)
    output.save("output.mp3")
    process = subprocess.Popen("start output.mp3", shell=True)
    processes.append(process)

output = [p.wait() for p in processes]

About mediapipe holistics

Mediapipe holistics tracking

About tensorflow keras

Tensorflow keras APIs

About Google Text To Speech (gtts)

Documentation of gtts