查詢結果分析
來源資料
相關文獻
- 以類神經網路輔助加馬刀立體定位放射手術治療計劃決策之初步研究
- Neural Network Procedures for Taguchi's Dynamic Problems
- A Fast and Efficient Competitive Learning Design Algorithm Based on Weight Vector Training in Transform Domain
- 專家系統振動訊號圖型判別之研究
- 反傳遞模糊類神經網路於流量推估之應用
- 類神經網路(Neural Networks)的種類及其在影像處理上的應用
- C++Fuzzy類神經網路物件導向發展系統之建立
- 臺灣汽保費率之估計--對數線性費率模式與類神經網路之比較
- 運用類神經網路於股價指數之套利--以日經225指數為例
- 使用類神經網路預估碳化鎢材料放電加工性能
頁籤選單縮合
題名 | 以類神經網路輔助加馬刀立體定位放射手術治療計劃決策之初步研究=Neural Network Assisted Gamma Knife Radiosurgery Treatment Planning: Basic Studies |
---|---|
作者姓名(中文) | 曾尹俊; 朱唯勤; 詹寶珠; 鍾文裕; 潘宏基; | 書刊名 | 中華醫學工程學刊 |
卷期 | 18:2 1998.06[民87.06] |
頁次 | 頁97-105 |
分類號 | 416.3 |
關鍵詞 | 加馬刀放射手術; 治療計劃; 類神經網路; Gamma knife radiosurgeny; Treatment planning; Neural network; |
語文 | 中文(Chinese) |
中文摘要 | 我們使用錯誤倒傳遞類神經網路演算法,訓練對象包括自行建造之32個假體病例 ,以及取樣自臺北榮總神經外科自民國 82 年 3 月起, 治療成功的 74 個實際病例,經處 理編碼等步驟,以病患之腫瘤病灶為輸入變數,治療計劃為輸出變數,做監督式學習。整個 系統建構在 Pentium 個人電腦上, 作業環境為 MS-DOS+Windows3.1, 程式設計語言採用 Borland C++4.5 之物件導向的觀念來開發。 實驗的結果顯示只有在使用假體做為學習範例 時,才能得到較佳的網路收斂情形。對於真實病例中可能出現的複雜病灶外形,仍無法達到 一個可以接受的程度。我們將針對此一研究做一檢討,布望找到設計上的缺失,以作為日後 改進的依據。 |
英文摘要 | Due to the extreme accuracy requirement of Gamma knife stereotactic radiosurgery, the treatment planning procedure has been a time consuming and enduring process for neurosurgeons to uphold. For this reason, we propose using artificial neural network to integrate traditional intricate diagnosing procedures to determine multi-focal shots, focal positions and weightings in an automatic or semi-automatic way. We use error-back -propagation neural network algorithm to train 32 phantom cases and 74 real cases acquired from the Neurological Institute of Veterans General Hospital, Taipei. After pre-processing procedure, tumor lesion data was encoded as input vector and the treatment result as the output vector of the artificial neural network. The program was constructed with an object-oriented programming concept that runs on a PC platform. Partial successes have been obtained for simple-shaped tumors, however, for complex-shaped and/or large tumors, further endeavors are required. Discussions on this are included. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。