سال انتشار: ۱۳۸۵

محل انتشار: سیزدهمین کنفرانس مهندسی پزشکی ایران

تعداد صفحات: ۶

نویسنده(ها):

Z Torbatian –
H Soltanian-Zadeh –

چکیده:

Magnetic Resonance Spectroscopic Imaging (MRSI) provides useful information about chemical properties of the tissue. In this paper, a new method was developed based on dyadic Wavelet transform to extract metabolite information of MRSI spectra like N-acetyl aspartate (NAA), Choline (Cho), and Creatine (Cr) appeared in specific locations of signal. After differentiation of brain signals from background based on signal energy, the wavelet coefficients were processed in order to detect peaks.Detection of positive maximum-negative minimum (modulo maxima) pairs in wavelet coefficients and considering predefined metabolite rangesin the signal lead to NAA, Cho, and Cr differentiation. It was tested and evaluated on simulated data and real data of 10 patients. The mean accuracy of the proposed method was 90% and 89% for simulated data and real data respectively.