Bioautomation

Volume 21, Issue 3
Bioinformatics
Research on Improved Fingerprint Image Compression and Texture Region Segmentation Algorithm	231-240
Kun Liu [ +/- abstract ][ full text ]
Targeted at the image segmentation and compression problems commonly found in fingerprint image processing, this paper puts forward an improved image segmentation algorithm based on grayscale normalization and an improved image compression algorithm based on sparse matrix transform. The proposed algorithms are proved effective through checking computations. The research results show that: the normalization, open operation and smoothing of the original fingerprint image bring about significant improvement to image segmentation effect, effective elimination of the noise and non-texture interference information in the fingerprint image, and strong guarantee of the real-time effect of fingerprint image processing through the calculation of time periods; the redundancy of non-texture information in the image is effectively reduced by converting the fingerprint image into sparse domains, and the blocking effect is suppressed by calculating the mean grayscale and low frequency information in the image. According to the image compression experiment, the proposed algorithm demonstrates excellent rate distortion feature. The compression process preserves most of the key information features of the original image, and realizes high the imaging quality.
Assessment of Data Reliability of Wireless Sensor Network for Bioinformatics	241-250
Ting Dong [ +/- abstract ][ full text ]
As a focal point of biotechnology, bioinformatics integrates knowledge from biology, mathematics, physics, chemistry, computer science and information science. It generally deals with genome informatics, protein structure and drug design. However, the data or information thus acquired from the main areas of bioinformatics may not be effective. Some researchers combined bioinformatics with wireless sensor network (WSN) into biosensor and other tools, and applied them to such areas as fermentation, environmental monitoring, food engineering, clinical medicine and military. In the combination, the WSN is used to collect data and information. The reliability of the WSN in bioinformatics is the prerequisite to effective utilization of information. It is greatly influenced by factors like quality, benefits, service, timeliness and stability, some of them are qualitative and some are quantitative. Hence, it is necessary to develop a method that can handle both qualitative and quantitative assessment of information. A viable option is the fuzzy linguistic method, especially 2-tuple linguistic model, which has been extensively used to cope with such issues. As a result, this paper introduces 2-tuple linguistic representation to assist experts in giving their opinions on different WSNs in bioinformatics that involve multiple factors. Moreover, the author proposes a novel way to determine attribute weights and uses the method to weigh the relative importance of different influencing factors which can be considered as attributes in the assessment of the WSN in bioinformatics. Finally, an illustrative example is given to provide a reasonable solution for the assessment.
Biomedical systems
On Use of Independent Component Analysis for Ocular Artifacts Reduction of Electroencephalogram and while Using Kurtosis as the Threshold	251-260
Kazi Aminul Islam, Gleb Tcheslavski [ +/- abstract ][ full text ]
Brain electrical activity commonly represented by the Electroencephalogram (EEG), due to its miniscule amplitude (on the order of a hundred microvolts), is often contaminated with various artifacts. Independent Component Analysis (ICA) may be a useful technique to reduce some artifacts prior analyzing EEG. Present report discusses use of kurtosis to determine the threshold for detecting the artifacts-related independent components. Kurtosis may represent how peaked or how flat the artifacts that affect a signal are compared to the normal behavior of the original signal. Two statistical approaches were used for the kurtosis-based threshold selection: the Z-score and the confidence interval. The independent components determined as artifact-related may be either set to zero for the greater artifact suppression or scaled down for the reduced effect on the artifact-free regions of EEG. Based on the observed results, we may conclude that the present technique may be used for ocular artifacts reduction in EEG.
Mathematical Contributions to the Study of Diabetes Mellitus	261-268
A. G. Shannon [ +/- abstract ][ full text ]
The purpose of the studies outlined in this paper is to describe some indicative and non-standard, but not exhaustive, quantitative contributions to the collection of diseases within the Diabetes Mellitus (DM) spectrum. While the paper has implications for bioinformatics related to DM, the paper is broader than that; it is more about DM-related bioprocesses illuminated by bioinformatics than about the bioinformatics per se. In effect it is an argument against being locked into one particular paradigm in the laudable study of this complicated set of diseases which increasingly dominate public health budgets, not to mention the lives of the patients with DM and their families.
The Application in Edge Detection of Medical Image Based on the Improved B-spline Wavelet Transform	269-278
Jinsong Zhang [ +/- abstract ][ full text ]
The edge information of a medical image is the basis of the analysis and processing of that medical image. The medical image contains the muscle, blood vessels and other interference information, which is large but leads the edge detection to become very difficult. So the research into an image edge detection algorithm has become one of the key technologies in image processing, and has important significance in the field of practical application. A new edge detection algorithm based on wavelet transform is proposed in this paper on the basis of regression analysis. Firstly, the signal is decomposed into approximate part and detailed part by the wavelet transform. Secondly, the wavelet coefficients of the detailed part are influenced by the threshold. Finally, the regression function of the signal is obtained by reconstructing wavelet coefficients. The experimental results show that the signal-noise ratio of the improved algorithm is clearly higher than the original, and it is helpful for improving the regression accuracy in terms of the wavelet transform. The algorithm can effectively remove noise, and the local feature of the signals is fully retained.

Sponsored by National Science Fund of Bulgaria, Grant No DNP 05-40/2016