A Lip Characteristic Parameter-based Identity Recognition Algorithm
- DOI
- 10.2991/iset-15.2015.37How to use a DOI?
- Keywords
- Interpolation, Characteristics extraction, Lip motion
- Abstract
Most studies in lip-based speaker identity recognition fail to make full use of the changing characteristics of lips and have limitations of heavy information processing work and low rate of successful recognition. A new method of lip characteristics extraction and speaker identification based on the integrative analysis of both static and dynamic lip characteristics is proposed for this sake. The first step of this method is to extract the speaker’s lip image set and mark critical points; then figure up the function curve of the speaker’s mouth-corner points and underlip’s lower midpoints, the function curve of the speaker’s left mouth-corner’s outer edge, the left highest points of upper lip’s top edge and the midpoints of upper lip edge according to Lagrange Interpolation; then use the two kinds of coefficients integratively as movement characteristic parameters to analyze the similarity level between the unidentified speaker’s lip characteristics extracted with that in the sample bank and the speaker can then identified. The results of simulation experiment prove that this method guarantees a high rate of successful recognition as well as lower complexity of computation and thus enjoys a good usability.
- Copyright
- © 2015, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Yingjie Meng AU - Wei Chen AU - LiXin Bai PY - 2015/03 DA - 2015/03 TI - A Lip Characteristic Parameter-based Identity Recognition Algorithm BT - Proceedings of the First International Conference on Information Science and Electronic Technology PB - Atlantis Press SP - 145 EP - 149 SN - 2352-538X UR - https://doi.org/10.2991/iset-15.2015.37 DO - 10.2991/iset-15.2015.37 ID - Meng2015/03 ER -