Skip to content

This project uses Keras to predict number signs from 0 to 5. OpenCV is used to connect to webcam and collect test/train data.

Notifications You must be signed in to change notification settings

lavsharmaa/ml-asl-sign-detection

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ML Sign Detection (Digits 0 to 9)

How to use it

  • Download or clone the project using git clone https://github.com/lavsharmaa/ml-asl-sign-detection or fork the repo.

Creating your own dataset

  • collect-data.py is used for creating your own dataset.
  • Run it using python collect-data.py it will show you a list of numbers from 0 to 9.
  • Using the number you can create dataset. For example, if you want to create dataset of number 0 press 0 on your keyboard and perform the sign it will start capturing it.
  • Similarly, you can do this for other numbers also and even for alphabets A to Z.
  • We followed 80:20 ratio. That means if there are 100 images for digit 0 we placed 80 images for training and 20 images for testing. Higher the number of images in training and testing more accurate your model will be in detecting the sign.
  • You can also download the dataset that we have created.

Training the dataset

  • After you have successfully generated your dataset.
  • Run python train.py to generate your model. It will take few mins and you will get your training accuracy.

Predicting the sign

  • Finally run python predict.py, perform your sign into the webcam and see the output.

Tools and Technologies

Food Corner is the web based project. Developed using

  • Python (3.7.4)
  • IDE (Jupyter)
  • Numpy (version 1.16.5)
  • cv2 (openCV) (version 3.4.2)
  • Keras (version 2.3.1)
  • Tensorflow (as keras uses tensorflow in backend and for image preprocessing) (version 2.0.0)

Team

About

This project uses Keras to predict number signs from 0 to 5. OpenCV is used to connect to webcam and collect test/train data.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages