/Smart-Glove

A final year project of my Bachelor in Electronics and Communication Engineering

Primary LanguageJupyter NotebookMIT LicenseMIT

Smart-Glove

Sign Language recognition system (model) developed using Random Forest Classifier, that translates the sign language alphabets and common words into text and sound.


Hardware look

Hardware Image


Overall system Information

  • Supervised machine learning.

  • Random Forest Algorithm

  • System recognizes gestures through the use of flex sensor, accelerometer and gyroscope.

  • All the programming and integration has been done in mac and for mac, so adjust the port for arduino as per device port number.

  • Arduino_code.ino is burned inside the Arduino Mega2560.

  • python_code whenever is executed it fetches the data from the serial port of the laptop then passes the data from the model for prediction of output as well as display and audio.

Note : Dataset is developed for the used system, so is not made public.


Data Visualizations of overall sensor data

Overall Sensor Data Visualization


Some correlation plots of different sensors values for different alphabets

  • Correlation plot of alphabet A

Correlation plot of alphabet A


  • Correlation plot of alphabet B

Correlation plot of alphabet B


  • Correlation plot of alphabet C

Correlation plot of alphabet C


  • Correlation plot of alphabet D

Correlation plot of alphabet D


  • Correlation plot of alphabet E

Correlation plot of alphabet E


  • Correlation plot of alphabet M

Correlation plot of alphabet M


  • Correlation plot of alphabet N

Correlation plot of alphabet N


  • Correlation plot of alphabet O

Correlation plot of alphabet O


  • Correlation plot of alphabet S

Correlation plot of alphabet S


  • Correlation plot of alphabet T

Correlation plot of alphabet T


  • Correlation plot of alphabet U

Correlation plot of alphabet U


  • Correlation plot of alphabet V

Correlation plot of alphabet V


Authors

LICENSE

MIT License