Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/4843
標題: 類神經網路與隱藏式馬可夫隨機領域模型於腦部核磁共振影像分割之研究
On the Segmentation of Brain MR Images using Neural Network and Hidden Markov Random Field Model
作者: 宋威廷
Sung, Wei-Ting
關鍵字: MR images;核磁共振造影;Neural Network;hidden Markov random field model;類神經網路;隱藏式馬可夫隨機領域模型
出版社: 通訊工程研究所
摘要: 
腦部組織的切割不僅對腦部核磁共振造影的分析是有價值的,在醫學臨床的診斷和疾病的預防上也是有廣大的應用。在此篇論文中,我們首先將腦部核磁共振造影經過一個擷取顱內組織的前處理,然後經由類神經網路獲得初始分割。然而類神經網路並沒有考慮到鄰近區域的資訊,這會造成類神經網路只有在雜訊程度較低的情況下才有較好的分割結果。所以我們在類神經網路的初始分割過後,再進行了隱藏式馬可夫隨機領域的分類。最後,我們結合了T1向量、T2向量和PD向量三種不同分類結果的資訊獲得最終的分類。

The segmentation of the various brain tissues is very valuable in the analysis of brain MR images, and it has a wide range of applications such as clinical analysis and visual inspection. In this thesis, we first use a pre-processing technique, called skull-stripping, and then we acquire an initial segmentation of a brain MR image through a neural network. Since the neural network does not consider the information of the neighborhood, the neural network can only work well on images with low levels of noise. So we use the hidden Markov random field to classify brain MR images after the initial segmentation of the neural network. Finally, we associate with three different segmentation results of T1-weighted, T2-weighted and PD-weighted to obtain the final segmentation.
URI: http://hdl.handle.net/11455/4843
Appears in Collections:通訊工程研究所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.