Details
Presenter(s)
![Shuai Li Headshot](https://confcats-catavault.s3.amazonaws.com/CATAVault/ieeecass/master/files/styles/cc_user_photo/s3/user-pictures/13964.jpg?h=16a003b7&itok=nzVyOhfz)
Display Name
Shuai Li
- Affiliation
-
AffiliationUniversity of Electronic Science and Technology of China
- Country
Abstract
This paper presents a deep Bidirectional Independently Recurrent Neural Network for the hand gesture recognition task. The Bidirectional Independently Recurrent Neural Network (Bi-IndRNN) is developed to better explore the temporal dependency by processing the inputs (and the following hidden states) from two directions. In addition, the temporal displacement is used to further enhance the input features with explicit description on the temporal movement. State-of-theart performance has been achieved on the DHG 14/28 dataset with classification accuracies of 93.15% and 91.13% for the 14 gesture classes case and the 28 gesture classes case, respectively.