Elmira Yekani Khoei
1, Reza Hassannejad
2*, Behzad Mozaffari Tazehkand
31 Faculty of Computer, College of Engineering, East Azerbaijan Science and Research Branch, Islamic Azad University, Tabriz, Iran
2 Faculty of Mechanical Engineering, University of Tabriz, Tabriz, Iran
3 Faculty of Electrical and Computer Engineering, University of Tabriz, Tabriz, Iran
Abstract
Introduction:
Body
sensor network is a key technology that is used for supervising the physiological
information from long distance that enables physicians to
predict, diagnose effectively the different conditions from long distance. These
networks include small sensors with the ability of sensing where there are some limitations
in calculating and energy.
Methods: In the present research, a new
compression method based on the analysis of principal components and
wavelet transform is used to increase the coherence. In the present method, the
first analysis of the main principles is used to find the principal components
of the data in order to increase the coherence for increasing the similarity
between the data and compression rate. Then, according to the ability of
wavelet transform, data are decomposed to different scales. In restoration
process of data only special parts are restored and some parts of the data that
include noise are omitted. By noise omission, the quality of the sent data
increases and good compression could be obtained.
Results: Pilates
practices were executed among twelve patients with various dysfunctions. The
results show 0.7210, 0.8898, 0.6548, 0.6765, 0.6009, 0.7435, 0.7651, 0.7623,
0.7736, 0.8596, 0.8856 and 0.7102 compression ratio in proposed method and 0.8256,
0.9315, 0.9340, 0.9509, 0.8998, 0.9556, 0.9732, 0.9580, 0.8046, 0.9448, 0.9573
and 0.9440 compression ratio for previous method (Tseng algorithm).
Conclusion: Comparing compression rate
and prediction errors with the available results shows the exactness of the proposed
method.