Subunit sign modeling framework for continuous sign language recognition

被引:16
作者
Elakkiya, R. [1 ]
Selvamani, K. [2 ]
机构
[1] SASTRA Univ, Sch Comp, Dept Comp Sci & Engn, Thanjavur, India
[2] Anna Univ, DCSE, Chennai, Tamil Nadu, India
关键词
Sign language recognition; Subunit sign modeling; Continuous sign recognition; Bayesian parallel hidden Markov model; SEGMENTATION; LEXICON;
D O I
10.1016/j.compeleceng.2019.02.012
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
A new framework named three subunit sign modeling is introduced for automatic sign language recognition. This works on continuous video sequences consisting of isolated words, signed sentences under different signer variations and illuminations. Three major issues of automatic sign language recognition is addressed namely: (i) importance of discriminative feature extraction and selection (ii) handling epenthesis movements and segmentation ambiguities (iii) automatic recognition of large vocabulary sign sentences and signer adaptation in a single subunit sign modeling framework. The proposed work has been evaluated and experimented subjectively and quantitatively with real-time signing videos gathered from different corpora and different sign languages. The results of the experiments have proven that the proposed subunit sign modeling framework remains scalable while increasing the sign vocabulary. Further, the approach is more reliable and efficient enough to adapt to real-time constraints and signer independence. (C) 2019 Elsevier Ltd. All rights reserved.
引用
收藏
页码:379 / 390
页数:12
相关论文
共 24 条
  • [1] Bauer B., 2001, INT GESTURE WORKSHOP, P64
  • [2] Christian Vogler, 2012, P 5 WORKSH REPR PROC
  • [3] Cooper H, 2017, SPRING SER CHALLENGE, P89, DOI 10.1007/978-3-319-57021-1_3
  • [4] Definition and recovery of kinematic features for recognition of American sign language movements
    Derpanis, Konstantinos G.
    Wildes, Richard P.
    Tsotsos, John K.
    [J]. IMAGE AND VISION COMPUTING, 2008, 26 (12) : 1650 - 1662
  • [5] Enhanced dynamic programming approach for subunit modelling to handle segmentation and recognition ambiguities in sign language
    Elakkiya, R.
    Selvamani, K.
    [J]. JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 2018, 117 : 246 - 255
  • [6] Extricating Manual and Non-Manual Features for Subunit Level Medical Sign Modelling in Automatic Sign Language Classification and Recognition
    Elakkiya, R.
    Selvamani, K.
    [J]. JOURNAL OF MEDICAL SYSTEMS, 2017, 41 (11)
  • [7] Large vocabulary sign language recognition based on fuzzy decision trees
    Fang, GL
    Gao, W
    Zhao, DB
    [J]. IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART A-SYSTEMS AND HUMANS, 2004, 34 (03): : 305 - 314
  • [8] Automatic skin segmentation and tracking in sign language recognition
    Han, J.
    Awad, G.
    Sutherland, A.
    [J]. IET COMPUTER VISION, 2009, 3 (01) : 24 - 35
  • [9] Boosted subunits: a framework for recognising sign language from videos
    Han, Junwei
    Awad, George
    Sutherland, Alistair
    [J]. IET IMAGE PROCESSING, 2013, 7 (01) : 70 - 80
  • [10] Lexicon-free fingerspelling recognition from video: Data, models, and signer adaptationI
    Kim, Taehwan
    Keane, Jonathan
    Wang, Weiran
    Tang, Hao
    Riggle, Jason
    Shakhnarovich, Gregory
    Brentari, Diane
    Livescu, Karen
    [J]. COMPUTER SPEECH AND LANGUAGE, 2017, 46 : 209 - 232