CHANG, RUEY FENG,张瑞峰,張瑞峰,HUANG, CHIUN SHENG,黄俊升,黃俊升,CHOU, YI HONG,周宜宏,周宜宏,CHANG, YEUN CHUNG,张允中,張允中,HSU, WEI WEN,徐位文,徐位文,SHEN, YI WEI,沈毅伟,沈毅偉,HUANG, YAN HUANG,黄彦皓,黃彥皓
申请号：
TW101124714
公开号：
TW201402074A
申请日:
2012.07.10
申请国别(地区):
TW
年份:
2014
代理人:
摘要：
This invention relates to an ultrasound imaging breast tumor detection system and its method. The invention is to acquire a plurality of three-dimensional breast ultrasound images and use a mean shift algorithm to cut the breast ultrasound images into a plurality of small areas, and acquire a mean grayscale value from each area, then the corresponding areas are categorized according to the grades of their mean grayscale values. The area with lowest grade is considered as a suspicious tumor area. The area with lowest grade is combined with at least one adjacent area having similar properties to form a completed area of suspicious tumor tissue. The tumor characteristics of the completed area of suspicious tumor tissue is analyzed to judge whether it is a tumor tissue or a non-tumor tissue. Since the invention uses areas as basic computing units, it is capable of quickly detecting tumor tissues from the breast ultrasound images.本發明有關於一種乳房超音波影像之腫瘤偵測系統及其方法,首先擷取複數張三維乳房超音波影像,利用一平均位移演算法以將乳房超音波影像切割成複數個小塊區域,從各區域之中分別擷取出一平均灰階值,依照各平均灰階值所屬的等級分類其對應的各區域,分類為最低等級之區域被視為一可疑腫瘤區域,再將最低等級之區域與其至少一相鄰且同質性相似之區域合併為一可疑腫瘤組織完整區域,分析可疑腫瘤組織完整區域之腫瘤特徵以判別為一腫瘤組織或一非腫瘤組織,則,係以區域為基本運算單元,將可以快速地從乳房超音波影像之中偵測出腫瘤組織。100‧‧‧腫瘤偵測系統11‧‧‧影像擷取模組111‧‧‧三維乳房超音波影像12‧‧‧影像切割模組120‧‧‧區域1200‧‧‧平均灰階值121‧‧‧區域122‧‧‧可疑腫瘤組織完整區域1220‧‧‧腫瘤特徵123‧‧‧區域124‧‧‧可疑腫瘤組織完整區域1240‧‧‧腫瘤特徵13‧‧‧平均灰階值擷取模組14‧‧‧區域分類模組15‧‧‧區域合併模組16‧‧‧特徵擷取分析模組17‧‧‧腫瘤標示模組171‧‧‧腫瘤地圖18‧‧‧使用者介面
