This code was originally written for a student project at the Technical University Munich. The goal was to train a state of the art (2017) deep CNN to classify RGB images showing finger spelling signs. We achieved an accuracy of 83.6% on a subset of speaker E from the ASL Finger Spelling Dataset with a DenseNet 50. The model was trained on speaker A to D as well as on images from RWTH Finger Spelling Database.
-
Notifications
You must be signed in to change notification settings - Fork 1
TheFloe1995/pytorch-ASL
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
No description or website provided.
Topics
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published