| Summary: | 碩士 === 國立臺灣大學 === 電機工程學系研究所 === 86 === A phenomenon can be found which abnormal joint sound arises from knee join
tdisorders during knee motion in the clinical diagnosis. There are some resear
cheswhich analyze the chondromalacia patella signals using mathematical model
bepublished. Nevertheless, as far as degenerative osteoarthritis and artifical
knee joint in vivo are concerned, there are practically no papers be published
.The hypothesis in the thesis is if the knee joint breaks down, abnormal vibra
tion signals will be generated within the knee joint. The main methods in t
he thesis include both autoregressive modeling, utilizingit to model the vibra
tion signals, and adaptive segmentation, dividing the vibrationsignals into se
gments. The power spectral ratio of dominant poles, root mean square, and intr
aclass distance are the feature parameters which appraise the state of the kne
e joint. We extract the spectral power ratio of dominant poles and root meansq
uare in the research of artifical knee joint in vivo, whereas the spectralpowe
r ratio of dominant poles and intraclass distance are used inthe research of d
egenerative osteoarthritis. Comparing the method utilizing vibration arthro
metry to measure the vibrationsignals arising from knee joint and the signal p
rocessing technique to analyze the vibration signals to detect the state of kn
ee joint with X-ray imaging, computertomography (CT) scan, magnetic reasonce i
maging (MRI), and arthroscopy, etc.. Wehave provided a noninvasive, cheap, and
simple clinical tool for diagnosingpatients with artificial knee joint and de
generative osteoarthritis.
|
|---|
