Static fingerspelling recognition based on boundary tracing algorithm and chain code

Citation

Abdullah, Junaidi and Dawod, Ahmad Yahya and Nordin, Md Jan (2018) Static fingerspelling recognition based on boundary tracing algorithm and chain code. In: ISMSI '18: Proceedings of the 2nd International Conference on Intelligent Systems, Metaheuristics & Swarm Intelligence. Association for Computing Machinery, pp. 104-109. ISBN 9781450364126

[img] Text
3206185.3206195.pdf - Published Version
Restricted to Repository staff only

Download (948kB)

Abstract

This paper presents a novel method for the detection and extraction of shape feature for fingerspelling recognition using boundary tracing and chain code. The method includes several steps such as conversion of RGB to YCbCr color space of an image and segmentation of skin pixel regions using thresholding method in order to construct binary images. Edge detection is applied and the location of candidate fingertips is estimated based on boundary tracing process and local extrema. The modified 2D chain code algorithm is then applied to the edge image to extract the fingerspelling shape feature and Support Vector Machine (SVM) is used for the classification task. The experimental findings show that the accuracy of the proposed method is 97.75% and 96.48% for alphabets and numbers, respectively.

Item Type: Book Section
Uncontrolled Keywords: Boundary value problems, chain code, shape feature, fingerspelling, Fingertips detection, hand tracing
Subjects: Q Science > QA Mathematics
Divisions: Faculty of Computing and Informatics (FCI)
Depositing User: Ms Suzilawati Abu Samah
Date Deposited: 20 Jan 2021 04:15
Last Modified: 20 Jan 2021 04:15
URII: http://shdl.mmu.edu.my/id/eprint/7328

Downloads

Downloads per month over past year

View ItemEdit (login required)