結合力學理論與基因演算法之模糊類神經網路應用於乳房腫瘤探測

Fuzzy Neural network Combining Theory of Mechanics and Genetic Fuzzy Neural network Combining Theory of Mechanics and Genetic Algorithms for Exploration of Breast Tumor

黃添康(Tian-Kang Huang)

指導教授 ： 顏炳郎

國立臺北科技大學/機電學院/自動化科技研究所/碩士(2009年)

https://doi.org/10.6841/NTUT.2009.00242

摘要

乳癌不僅已連續五年居於國內女性惡性腫瘤死因之第二位。在美國的婦女中，每八人就有一人罹患乳癌，而在已開發國家中，乳癌甚至是女性主要之死亡原因。就乳癌早期診斷而言，乳癌攝影是目前最為廣泛使用的方法。然而乳房攝影對於乳癌的高偽陽性，已導致了相當高比例之不必要的組織切片檢查。一般而言，組織切片檢查是一個既昂貴又令人有挫折感的程序。近年來許多研究都證實使用乳房超音波影像可以有效的降低組織切片的數目。可是這些研究全憑醫師的專業知識和個人經驗做出判斷，造成結果主觀，經驗不同的醫師診斷常有不同的結果產生。另外在臨床上區分乳房腫瘤良性或惡性與腫瘤和軟組織之硬度比也有其關連性，通常越硬的腫瘤惡性的機率越高。本文以生物力學的角度分析人體組織與腫瘤組織之特性，製作出乳房腫瘤力學模型，並經由橫向探測實驗找出具有描述腫瘤特性之力量特徵參數群，做為本研究判斷腫瘤硬度比的資料來源，然而對於判斷腫瘤軟硬度的此種行為，並沒有辦法以數學模型來描述，所以可以透過模糊類神經網路(Fuzzy Neural Network)對乳房腫瘤力學模型進行模糊建模，以取代數學模型描述，且模糊建模經由輸出回授，反覆進行系統誤差的修正，以得到預期的腫瘤與軟組織硬度比。

關鍵字

模糊類神經網路 ； 電腦輔助診斷 ； 超音波影像

並列摘要

Breast cancer has been ranked the fourth leading cause of cancer deaths for females in Taiwan for the past five years in a row. It has also been reported that one of every eight women in the USA was affected by the breast cancer and the breast cancer had become the first cause of death for the female populations in the developed countries. While mammography has been the most widely used approach for early detection of breast cancers, the high false positive rate for the breast lesions using mammography has led to a high percentage of unnecessary biopsy referral, which is an expensive and disconcerting procedure. Recently, many studies have shown that it has a very high potential to use the breast sonography to reduce the number of biopsies. Nevertheless, these studies were usually made by highly experienced medical doctors and the sonographic features suggested by them were very dependent on the interpretation of the ultrasound images. First a membership function with adaptive spans under the fuzzy orthogonal condition is established to define the fuzzy set. Then using the genetic algorithms, the modulated membership factors(MMF) of membership function with adaptive span are learned automatically, so that the fuzzy inference system of automatic weighing is constructed. In this study, the fuzzy neural network is divided to premise and consequence, to establish the fuzzy modeling.

並列關鍵字

Fuzzy neural network ； Computer-aided diagnosis ； Ultrasound images

參考文獻

[1] 王信凱，超音波彈性影像 Ultrasound Elastography，中

[2].Abbas Samani, Judit Zubovits, Donald Plewes ”Elastic moduli of normal and pathological human vreast tissues: an inversion-technique-based investigation of 169 samples”.Phys.Med Biol 52(2007)156-1579.

[3]. Rafayah Mousa,Qutaishat Munib” Breast cancer diagnosis system based on wavelet analysis and fuzzy-neural”,Expert System with Application 28(2005) 713-723.

[6]. Chen DR, Chang RF, Chen WM, Moon WK.”Computer aided diagnosis for 3-dimensional breast ultrasonography”. Archives of Surgery 2003; 138(3): 296- 302.

[7]. Ruey-Feng Chang, Wen-Jie Wu,Woo kyung Moon,and Dar-Ren Chen,”Autom- atic ultrasound segmentation and morphology based diagnosis of solid breast tumors”. Breast Cancer Research and Treatment(2005) 89:179-185.