機器手臂應用於乳房腫瘤之偵測

Abstract

乳癌是為目前女性發生率最高的癌症，但早期的檢測、治療可以增加乳癌的治癒率。乳房觸診為常用來進行初步診斷的方法之一，利用觸診的方式也較為主觀，且有不能量化等缺點。為了提升準確度，本研究設計一個乳房腫瘤的輔助診斷系統，可以輔助醫生來進行觸診提高準確度，並提供量化且客觀的資訊。 系統中為利用自動化的器械，透過機器手臂及力量感測器來對乳房進行檢查，來診斷乳房是否有腫瘤，並探測乳房腫瘤的位置。利用力量感測器感測腫瘤與正常乳房組織的差異性，來輔助醫師提高醫師在判斷乳房腫瘤的準確率。 本研究在利用探測機制過後的結果：陽性為84.6%、偽陽性為15.4%、偽陰性為4.16%。而整體的探測成功率為95.8%，而後判斷腫瘤的位置，其均方誤差為2.44，相對誤差4.32%。

Breast cancer is currently the highest incidence of female cancers, but early detection and treatment of breast cancer can increase the cure rate. Breast palpation is used to conduct a preliminary diagnosis method. Using palpation way is also more subjective and can not be quantified and other shortcomings. To improve detecting accuracy, this study design a breast cancer diagnosis system that can assist doctors improve accuracy, and provide quantitative and objective information. System for the use of automated equipment, through the robotic machine and sensors to breast examination to diagnose whether breast tumors, and breast tumors position. The use of force sensor sensing a tumor and normal breast tissue differences, to assist physicians in improving the accuracy of physician judgment breast tumors. In this study, the use of detection mechanisms after the result: positive 84.6%, 15.4% false positives, false negative 4.16%. The overall success rate of 95.8% to detect, and then determine the location of the tumor, the mean square error of 2.44, the relative error of 4.32%.