Skip to content

Keras and Tensorflow based CNN to identify images as one of 26 classes detailing the signs in the American Sign Language alphabet.

Notifications You must be signed in to change notification settings

otter-lights/asl-recognizer

Repository files navigation

asl-recognizer

Keras and Tensorflow based CNN to identify images as one of 26 classes detailing the signs in the American Sign Language alphabet.

The model used is not currently in this repo because the file is too large, it can be replicated by running the final script and naming the created file "200p-model.h5" which is then called when classifying images.

Looking back at the data (from https://www.kaggle.com/grassknoted/asl-alphabet) the images are not of the best quality as some signs are positioned incorrectly or have poor lighting. This was more of a proof-of-concept than anything else and completed for a software engineering project for school (@CWcomptech on Twitter)

About

Keras and Tensorflow based CNN to identify images as one of 26 classes detailing the signs in the American Sign Language alphabet.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published