Scatter correction of PET based on non-linear statistical estimate method / 生物医学工程学杂志

Scatter coincident events are the important factor that affects the quality of positron emission tomography (PET) images. In this paper, according to the characters of projection data, a scatter correction method which uses maximum likelihood expectation maximization (MLEM) algorithm based on poisson model is proposed. We compared the sinograms and reconstructed images corrected by MLEM algorithm and deconvolution method respectively. The results indicate that the algorithm proposed in this paper increases the contrast of images while correcting scatter. It is better than the traditional method.

Hidden Markov model used in protein sequence analysis / 生物医学工程学杂志

Hidden Markov model (HMM) used in the research of protein is a new field of bioinformatics. Nowadays large amount of data about protein sequences and structures have been obtained. Traditional methods of protein analysis are no longer used. Biologists have updated their research methods with computer technology and statistics, which can deal with large amount of data. HMM can be used to distinguish protein sequence with the same characteristics. A family of protein from SCOP database was selected, through which a HMM model representing the family was obtained, and then the model was utilized to analyze protein sequences. Results indicate that HMM can express particular family of protein, and recognize the given protein sequences of the family from many sequences.

effects of various degrees of coronary stenosis on Lyapunov exponents of ECG and R-R interval time series / 中国医学物理学杂志

Purpose:It has been shown that the heart is a chaotic oscillator. So it is appropriiate to use the Lyapunov exponent, an important parameter to identify the nature of non-linear dynamical systems, for identifying the state of human heart. Methods:Preliminary results are obtained in this paper using Wolf's algorithm for 8 normal and 107 abnormal ECG recordings. Results:Significant differences are found between the Lyapunov exponents of normal ECG and ECG with obvious coronary stenosis (OCS), but there is no significant difference between the Lyapunov exponents of normal ECG and ECG with mild coronary stenosis (MCS);Significant differences are also found between the Lyapunov exponents of R-R interval series of normal ECG、ECG with MCS and ECG with OCS. Conclusions:It is apparent that the R-R interval series can give us more messages about human heart, and the Lyapunov exponents of ECG and R-R interval series are the appropriate parameters for the identification of the physiological states of human heart. It is possible to use Lyapunov exponent for early diagnosis of Coronary Heart Disease.