Skip to content

PramodBobade/Sign-Language-Recognition-using-modified-AlexNet

Repository files navigation

Sign-Language-Recognition-using-modified-AlexNet

Alphanumeric ASL signs recognition by using Deep Learning's AlexNet Model.

Dataset

Dataset is taken from Massey University's Gesture database 2012. url- http://www.massey.ac.nz/~albarcza/gesture_dataset2012.html

Requirements

Following are all the packages that need to be installed before running this project.

Installation

numpy

os

cv2

random

tensorflow

keras

pandas

All the above packages can be installed using the following command(Replace packname with appropriate package name, without quotes):

pip install "packname"

Usage

For training the images has to be imported from the hands directory. Run the realtraining.ipynb file for training the model and this will eventually save the weights to finalweights.h5 in the present working directory.

For prediction, save the images file to predict in the present working directory. Now run the prediction.ipynb file which will eventually uses the weights from fimalweights.h5 file and prints the output of the prediction.

About

Alphanumeric ASL signs recognition by using Deep Learning's AlexNet Model.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors