Assistive technology for relieving communication lumber between hearing/speech impaired and hearing people

The Journal of Engineering

  ^{1    1    2    3    3    4    5    5    5} 

[1] Faculty (Kulliyyah) of Engineering, International Islamic University Malaysia, Jl, Gombak 53100, Kuala Lumpur, Malaysia;Faculty (Kulliyyah) of Engineering, International Islamic University Malaysia, Jl, Gombak 53100, Kuala Lumpur, Malaysia;School of Engineering and Advanced Technology, Massey University, New Zealand, Private Bag 11222, Palmerston North 4442, New Zealand;School of Engineering and Advanced Technology, Massey University, New Zealand, Private Bag 11222, Palmerston North 4442, New Zealand;School of Engineering and Advanced Technology, Massey University, New Zealand, Private Bag 11222, Palmerston North 4442, New Zealand;Centre of Technology, RMIT University Vietnam, 702 Nguyen Van Linh Blvd, Ho Chi Minh City, HCMC, Vietnam;School of Engineering, Monash University Malaysia, Jl Lagoon Selatan, 46150, Selangor Darul Ehsan, Malaysia;

关键词: sign language recognition;    hearing;    handicapped aids;    assisted living;    language translation;    feature extraction;    hidden Markov models;    Linux;    virtual machines;    assistive technology;    hearing impaired people;    speech impaired people;    automatic sign language translator;    speech impaired communities;    hearing impaired communities;    system architecture;    feature extraction;    feature recognition;    static signs;    dynamic signs;    hidden Markov model;    segmentation signature;    Windows7 operating systems;    LINUX Fedora 16 operating systems;    VMware workstation;    Malaysian sign language;   

DOI  :  10.1049/joe.2014.0039

来源: publisher

【 摘 要 】

This study proposes an automatic sign language translator, which is developed as assistive technology to help the hearing/speech impaired communities to communicate with the rest of the world. The system architecture, which includes feature extraction and recognition stages is described in detail. The signs are classified into two types: static and dynamic. Various types of sign features are presented and analysed. Recognition stage considers the hidden Markov model and segmentation signature. Real-time implementation of the system with the use of Windows7 and LINUX Fedora 16 operating systems with VMware workstation is presented in detail. The system has been successfully tested on Malaysian sign language.

【 授权许可】

CC BY