|Title:||Vision-based sign language recognition|
|Subject:||Optical pattern recognition.|
Human locomotion -- Analysis -- Data processing.
Hong Kong Polytechnic University -- Dissertations
|Department:||Department of Computing|
|Pages:||viii, 59 leaves : ill. ; 30 cm.|
|Abstract:||Sign language recognition is an important subfield of Human Computer Interaction (HCI), which enables people communicating with hearing impaired people conveniently. Many different methods are proposed and apply to the recognition problem previously including several dimensionality reducing methods and classification methods. However due to the complexity of human signs, it has never been an easy task. In this dissertation, different hand shapes of human signers are tracked and segmented from dynamic image sequences with analysis of RGB data and depth data captured by a Kinect sensor. A sparse autoencoder is trained to reconstruct hand shape images into high-level features. The learned features were then fed into a multinomial logistic regression model and trained it to classify signs of digit from 0-9 of Chinese Sign Language (CSL). A total of 4000 hand shape images are collected and used during the experiments and produced reasonable results. The performance of entire system is good to achieve almost realtime sign recognition after the model being trained.|
Files in This Item:
|b27522660.pdf||For All Users (off-campus access for PolyU Staff & Students only)||5.05 MB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: