The final accuracy of the model, based on 8 study subjects and 9,400 images, is 85% . The usage of CNNs is apromising approach to interpret sign languages, and transfer learning can be used to achieve high testing accuracy despite using a small training dataset . We describe the implementation details of our model to interpretsigns as a user-friendly web application . The implementation details are available in a web application to interpret signs of the Swedish Sign Language (SSL) handalphabet, which is based on a pre-trained InceptionV3 network and the usage of the mini-batch gradient descent optimizational algorithm . We rely on transfer learning during the pre-training of the . modeland its data

Author(s) : Gustaf Halvardsson, Johanna Peterson, C├ęsar Soto-Valero, Benoit Baudry

Links : PDF - Abstract

Code :

Keywords : learning - transfer - sign - interpret - application -

Leave a Reply

Your email address will not be published. Required fields are marked *