TW201126260A - Quantification method of the feature of a tumor and an imaging method of the same - Google Patents

Quantification method of the feature of a tumor and an imaging method of the same Download PDF

Info

Publication number
TW201126260A
TW201126260A TW99101930A TW99101930A TW201126260A TW 201126260 A TW201126260 A TW 201126260A TW 99101930 A TW99101930 A TW 99101930A TW 99101930 A TW99101930 A TW 99101930A TW 201126260 A TW201126260 A TW 201126260A
Authority
TW
Taiwan
Prior art keywords
tumor
value
gray
region
contour
Prior art date
Application number
TW99101930A
Other languages
Chinese (zh)
Other versions
TWI474284B (en
Inventor
King-Jen Chang
Wen-Hwa Chen
Argon Chen
Chiung-Nein Chen
Ming-Chih Ho
Hao-Chih Tai
Ming-Hsun Wu
Po-Wei Tsai
Chung-Wei Liu
Hsin-Jung Wu
Original Assignee
Amcad Biomed Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Amcad Biomed Corp filed Critical Amcad Biomed Corp
Priority to TW99101930A priority Critical patent/TWI474284B/en
Publication of TW201126260A publication Critical patent/TW201126260A/en
Application granted granted Critical
Publication of TWI474284B publication Critical patent/TWI474284B/en

Links

Landscapes

  • Image Analysis (AREA)
  • Image Processing (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

A quantification method and an imaging method are disclosed, capable of quantifying the margin feature, the cysts feature, the calcifications feature, the echoic feature and the heterogenesis feature of a tumor, and capable of imaging the margin feature, the cysts feature, the calcifications feature and the heterogenesis feature of a tumor. The quantification method and the imaging method calculate the moving variance of the gray scale of each of the pixel points based on the gradient value of the gray scale of these pixel points. Then, depending on the purpose of the quantification method or the imaging method, the maximum value, the minimum value, the mean value, and the standard deviation of the moving variance of the gray scale of these pixel points are calculated, respectively. At final, with the definition of the threshold value and the imaging rule, the above features of the tumor are quantified or imaged.

Description

201126260 六、發明說明： 【發明所屬之技術領域】 本發明係關於一種腫瘤特徵的量化方法及影像化方 法’尤指一種適用於將腫瘤所具有之邊緣特徵、囊腫特徵、 鈣化特徵'迴音性特徵及異質化特徵量化的量化方法及將 這些腫瘤特徵影像化之影像化方法。 【先前技術】 近年來，由於醫用超音波影像技術不論在影像的解析 度上及衫像數位化上均有顯著的進步，所以醫用超音波影 像技術除了被應用在胎兒成長狀況的監控之外，也逐步地 被應用在各種類型之腫瘤狀態的判斷上，例如甲狀腺腫 f 〇而且，由於醫用超音波影像技術之非侵入成像特性， 醫師也逐漸藉由醫用超音波影像技術的協助來判斷腫瘤性 貝及S乎估後續的處置方式。 而醬師攸腫瘤的超音波影像中判斷腫瘤性質的第一個 步驟是辨別出腫瘤的輪廓，以定義出腫瘤内部區域及腫瘤 外部區域。接著，醫師才能從對應於腫瘤内部區域之超音 U象的刀中辨別出腫瘤的各項特徵如邊緣特徵、囊 腫特徵、弼化特徵、迴音性特徵及異質化㈣等，做為其 判斷腫瘤質的參考。但是目前的醫用超音波影像系統 僅允。午醫師以肉眼觀察出他所認為之腫瘤輪廓再配合手 寫輸入裝置將此腫瘤輪廟輸人至腫瘤的超音波影像上。但 是’ S是此程序，就存在許多不可靠之處。 201126260 因為，目前的方式需完全倚賴醫師的主觀感受及經驗，甚 至醫師當時的精神狀態，所以針對同—腫瘤的超音波影 像’不同醫師所輸入的腫瘤輪廓並不相同，如圖ia所示二 :至，即便是同一位醫師，在不同時間針對同一腫瘤的超 音波影像所輸入的腫瘤輪廓也不盡相同。 然後’醫師才能藉由此腫瘤輪廓的協助，以肉眼觀察 的方式識別出腫瘤内部區域是否具有各項特徵及各項特徵 佔整個腫瘤内部區域的比率。最後，t師再藉由所蒐集到 的各項結果，士口某項特徵的分佈或某項特徵佔整個腫瘤内 :區域的比例等’ #1!斷出此腫瘤之性質。也就是說，目前 藉由腫瘤的超音波影像判斷腫瘤性質的程序並無一種客觀 的機制存在，導致誤判腫瘤性質的案例時有發生造成此 藉由醫用超音波影像技術判斷腫瘤性質的技術仍無法被醫 界及社會大眾信賴。 此外，雖然在影像辨識的領域（如車牌辨識）中已存在 數種影像邊緣辨識方式，如snake演算法。但是，此如此^演 算法在一開始仍須仰賴使用者輸入一起始邊緣（即醫師仍 須手動輸入一腫瘤的大略輪廓）至此演算法中，此snake演算 法才能開始後續的演算程序。況且，由於此5蠢演算法的 本身特性，其比較適合應用在影像邊緣明顯的案例中，否 則演算出的結果往往與實際邊緣差距極大。可是，腫瘤的 邊緣一般來說並不明顯，所以即使將此snake演算法應用在 腫瘤的超音波影像上，其所得出的腫瘤輪廓往往與腫瘤實 際輪廓仍存有一段差距，如圖丨B所示。況且，為求演算出 201126260 的輪廓比較接近腫瘤實際輪廓並縮短演算所需時間，醫師 還疋需要仔細地輸入一個與實際腫瘤輪廓不會相差太遠的 起始邊緣，結果還是沒有減輕醫師太多的負擔。除此之外， 由於腫瘤的超音波影像係為一種灰階影像，而腫瘤所具有 的各項特徵（如邊緣特徵、囊腫特徵、鈣化特徵 '迴音性特 徵及異g化特徵等）往往在此灰階影像中僅為某些像素點 所具有之灰階梯度值的些微改變，對於醫師的肉眼來說， 这些特徵並不容易辨識出來，醫師僅能憑著「感覺」來判 斷这些特徵是否存在’進而造成腫瘤性質的判斷僅能基於 醫師的主觀感《’而無法基於_實而精確地判斷。 因此，業界需要一種腫瘤特徵的量化方法及對應之影 像化方法’尤指-種適用於將腫瘤所具有之邊緣特徵囊 腫特徵、職特徵、迴音性特徵及異質化特徵量化的量化 方法及將這些腫瘤特徵影像化之影像化方法。 【發明内容】 本發明之主要目的係、在提供-種腫瘤特徵的量化方 法，俾能將腫瘤所1右沾々& ' /、有的各種特徵，如邊緣特徵、 徵、約化特徵、迴音性特^ ^ ^ ^ ^ ^ ^ ^ 評估之用。 寸R及異質化特徵罝化，以供醫師 本發明之另一目的私+ ,,的仏在k供一種腫瘤特徵的影像化方 法’俾此將腫瘤所具有的 n s H 的各種特徵，如邊緣特徵、囊腫特 貝化特徵影像化，以供醫師評估之用。 6 201126260 為達成上述目的，本發明之腫瘤邊緣特徵的量化方 法’係應用於一由複數個像素點組合而成並至少顯示一腫 瘤的灰階影像，包括下列步驟：（A)從此灰階影像擷取出一 腫瘤輪廓及一腫瘤輪廓環形區域，且此腫瘤輪廓係位於此 腫瘤輪廓環形區域内；（B)將此腫瘤輪廓重疊顯示於此灰階 影像上’以在此灰階影像上定義出一腫瘤内部區域及一腫 瘤外部區域；（C)操取此腫瘤輪廓環形區域的一重心點，定 義一從此重心點向外延伸並通過此腫瘤輪廓環形區域的剖 面線，及提供一位於此剖面線上並位於此腫瘤輪廓環形區 域内之量測線段；（D)計算出位於此量測線段上之此等像素 點所分別具有之灰階移動變異值；以及（E)依據位於此量測 線段上之每一此等像素點所分別具有之灰階移動變異值， 將位於此剖面線上之腫瘤邊緣特徵量化。 為達成上述目的’本發明之腫瘤邊緣特徵的影像化方 法，係應用於一由複數個像素點組合而成並至少顯示一腫 瘤的灰階影像，包括下列步驟：（A)從此灰階影像擷取出一 腫瘤輪廓及一腫瘤輪廓環形區域，且此腫瘤輪廓係位於此 腫瘤輪廓環形區域内；（B)將此腫瘤輪廓重疊顯示於此灰p皆 影像上，以在此灰階影像上定義出一腫瘤内部區域及一腫 瘤外部區域；（C)掘取此腫瘤輪廊環形區域的一重心點，定 義一從此重心點向外延伸並通過此腫瘤輪廓環形區域的剖 面線，及提供一位於此剖面線上並位於此腫瘤輪廓環形區 域内之量測線段；（D)計算出位於此量測線段上之此等像素 點所分別具有之灰階移動變異值；以及（E)依據位於此量測 201126260 線段上之此等像素點所分別具有之灰階移動變異值，定義 出一邊緣成像上界及一邊緣成像下界，且配合一彩虹色階 將位於此剖面線上之腫瘤邊緣特徵影像化。 為達成上述目的’本發明之腫瘤囊腫特徵的量化方 法’係應用於一由複數個像素點組合而成並至少顯示一腫 瘤的灰階影像，包括下列步驟：（A)從此灰階影像擷取出一 腫瘤輪廓及一腫瘤輪廓環形區域，且此腫瘤輪廓係位於此 腫瘤輪廓環形區域内；（B)將此腫瘤輪廓重疊顯示於此灰階 影像上，以在此灰階影像上定義出一腫瘤内部區域及一腫 瘤外部區域；（C)藉由位於此腫瘤内部區域内之此等像素點 所分別具有之灰階梯度值，計算出位於此腫瘤内部區域内 之此等像素點所具有之灰階梯度值的最小值及灰階梯度值 的標準差；以及（D)依據位於此腫瘤内部區域内之此等像素 點所具有之灰階梯度值的最小值及灰階梯度值的標準差， 將位於此腫瘤内部區域内之囊腫特徵量化。 為達成上述目的’本發明之腫瘤囊腫特徵的影像化方 法’係應用於一由複數個像素點組合而成並至少顯示一腫 瘤的灰階影像，包括下列步驟：（A)從此灰階影像擷取出一 腫瘤輪廓及一腫瘤輪廓環形區域，且此腫瘤輪廓係位於此 腫瘤輪廓環形區域内；（B)將此腫瘤輪廓重疊顯示於此灰階 〜像上，以在此灰階影像上定義出一腫瘤内部區域及一腫 瘤外部區域；（C)藉由位於此腫瘤内部區域内之此等像素點 所刀別具有之灰階梯度值，計算出位於此腫瘤内部區域内 之此等像素點所具有之灰階梯度值的最小值及灰階梯度值 8 201126260 的標準差；以及（D)依據位於此腫瘤内部區域内之此等像素 點所具有之灰階梯度值的最小值及灰階梯度值的標準差， 定義出一囊腫成像上界及一囊腫成像下界，以將位於此腫 瘤内部區域内之腫瘤囊腫特徵影像化。 為達成上述目的，本發明之腫瘤妈化特徵的量化方 法’係應用於一由複數個像素點組合而成並至少顯示一腫 瘤的灰階影像，包括下列步驟：（A)從此灰階影像擷取出一 腫瘤輪廓及一腫瘤輪廓環形區域，且此腫瘤輪廓係位於此 腫瘤輪廓環形區域内；（B)將此腫瘤輪廓重疊顯示於此灰階 影像上，以在此灰階影像上定義出一腫瘤内部區域及一腫 瘤外部區域；（C)藉由位於此腫瘤内部區域内之此等像素點 所分別具有之灰階梯度值’計算出位於此腫瘤内部區域内 之此等像素點所具有之灰階梯度值的最小值及灰階梯度值 的&準差；（D)依據位於此腫瘤内部區域内之此等像素點所 具有之灰階梯度值的最小值及灰階梯度值的標準差，從此 灰階影像中擷取出一位於此腫瘤内部區域内之囊腫區域； (E)藉由位於此腫瘤内部區域内但位於此囊腫區域之外之此 等像素點所分別具有之灰階梯度值，計算出位於此腫瘤内 部區域内但位於此囊腫區域之外之此等像素點所具有之灰 階梯度值的最大值、灰階梯度值的標準差及灰階梯度值的 平均值’以及（F)依據位於此腫瘤内部區域内但位於此囊腫 區域之外之此等像素點所具有之灰階梯度值的最大值、灰 階梯度值的標準差及灰階梯度值的平均值，將位於此腫瘤 内部區域内之腫瘤鈣化特徵量化。 201126260 為達成上述目的，本發明之腫瘤鈣化特徵的影像化方 法’係應用於一由複數個像素點組合而成並至少顯示一腫 瘤的灰階影像，包括下列步驟：（A)從此灰階影像擷取出一 腫瘤輪廓及一腫瘤輪廓環形區域，且此腫瘤輪廓係位於此 腫瘤輪廓環形區域内；（B)將此腫瘤輪廓重疊顯示於此灰階 影像上，以在此灰階影像上定義出一腫瘤内部區域及一腫 瘤外部區域；（C)藉由位於此腫瘤内部區域内之此等像素點 所分別具有之灰階梯度值，計算出位於此腫瘤内部區域内 之此等像素點所具有之灰階梯度值的最小值及灰階梯度值 的標準差；（D)依據位於此腫瘤内部區域内之此等像素點所 具有之灰階梯度值的最小值及灰階梯度值的標準差，從此 灰階影像中擷取出一位於此腫瘤内部區域内之囊腫區域； (E)藉由位於此腫瘤内部區域内但位於此囊腫區域之外之此 等像素點所分別具有之灰階梯度值，計算出位於此腫瘤内 部區域内但位於此囊腫區域之外之此等像素點所具有之灰 階梯度值的最大值、灰階梯度值的標準差及灰階梯度值的 平均值；以及（F)依據位於此腫瘤内部區域内但位於此囊腫 區域之外之此等像素點所具有之灰階梯度值的最大值、灰 階梯度值的標準差及灰階梯度值的平均值，定義出一的化 成像上界及一鈣化成像下界’以將位於此腫瘤内部區域内 之腫瘤鈣化特徵影像化。 為達成上述目的’本發明之腫瘤迴音性特徵的量化方 法’係應用於一由複數個像素點組合而成並至少顯示一腫 瘤的灰階影像，包括下列步驟：（八）從此灰階影像擷取出一 10 201126260 腫瘤輪廓及一腫瘤輪廓環形區域，且此腫瘤輪廓係位於此 腫瘤輪廓環形區域内；（B)將此腫瘤輪廓重疊顯示於此灰階 影像上’以在此夜階影像上定義出一腫瘤内部區域及一腫 瘤外部區域；（C)藉由位於此腫瘤内部區域内之此等像素點 所分別具有之灰階梯度值，計算出位於此腫瘤内部區域内 之此等像素點所具有之灰階梯度值的平均值；於此腫瘤 外部區域選取一基準區域，藉由位於此基準區域内之此等 像素點所分別具有之灰階梯度值，計算出位於此基準區域 内之此等像素點所具有之灰階梯度值的平均值；以及（E)依 據位於此腫瘤内部區域内之此等像素點所具有之灰階梯度 值的平均值及位於此基準區域内之此等像素點所具有之灰 階梯度值的平均值，將此腫瘤之迴音性特徵量化。 為達成上述自的，本發明之腫瘤異質化特徵的量化方 法，係應用於一由複數個像素點組合而成並至少顯示一腫 瘤的灰階影像，包括下列步驟：（A)從此灰階影像擷取出一 腫瘤輪廓及一腫瘤輪廓環形區域’且此腫瘤輪廓係位於此 腫瘤輪廓環形區域内；（B)將此腫瘤輪廓重疊顯示於此灰階 影像上，以在此灰階影像上定義出一腫瘤内部區域及一腫 瘤外部區域；（C)將位於此腫瘤内部區域内之此等像素點分 別定義為複數個參考遮罩’且每一此等參考遮罩係包含一 基準像素點與複數個相鄰於此基準像素點之像素點；計 算出每-此等參考遮罩所分別具有之參考遮罩灰階梯度值 局部平均及參考遮罩灰階梯度值局部變異；（ε)計算出每— 此等參考遮罩所分別具有之參考料灰㈣度值局部平均 201126260 之父異$考遮罩灰階梯度值局部變異之平均以及表考遮 罩灰階梯度值局部變異之變異；以及（F)藉由每一此等參考 遮罩所分別具有之$ ,卜, 至少一選自於一由參考遮罩灰階梯度值 局。p平均之變異、參考遮罩灰階梯度值局部變異之平均以 遮罩灰&梯度值局部變異之變異所構成之群組 =母1^等參考遮罩所分別具有之異質化指標值，將此 腫瘤之異質化特徵量化。 為=上述㈣，本發明之腫瘤異質化特徵的影像化 賭癌沾二：用於一由複數個像素點組合而成並至少顯示--腫瘤2衫像’包括下列步驟：⑷從此灰階影像操取出 及：腫瘤輪廓環形區域，且此腫瘤輪廓係位於 階影像上，以在此二重疊顯示於此灰 腫瘤外部區域;⑹將位於此腫瘤内部區域域及一 分別定義為複數個參考遮罩，且每-此等參考== 素點與複數個相鄰於此基準像素點之像素點二 ，3於每一此等參考遮罩之此等像素點所分別具有之 灰階梯度值，計算出每—此等參考 遮罩灰階梯度值變異；（Ε)藉由每7 : L有之參考 有之參考遮罩灰階梯度值變異’計算出此等參考 階梯度值變異;(F)藉由每-此等參考遮罩 之參考遮罩灰階梯度值變異及此等參考遮罩所 '、有之平均遮罩灰階梯度值變異， 罩所分別呈有之里質仆…s… 出母—此等參考遮 L有之異貝化g值，(G)藉由每 201126260 所分別具有之異質化指標值’計算出此等參考遮罩所具有 之異質化指標值的最大值、異質化指標值的最小值、異質 化指標值的平均值及異質化指標值的標準差；以及（H)依據 此等參考遮罩所具有之異質化指標值的最大值、異質化指 標值的最小值、異質化指標值的平均值及異質化指標值的 標準差’定義出一異質化成像上界及一異質化成像下界， 且配合一彩虹色階將位於此腫瘤内部區域内之腫瘤異質化 特徵影像化。 本發明提供一種電腦可讀取記錄媒體，係儲存有用以 使一電腦系統執行一腫瘤邊緣特徵之量化方法的程式，此 腫瘤邊緣特徵之量化方法係應用於一由複數個像素點组合 而成並至少顯示一腫瘤的灰階影像，包括下列步驟：（A)從 此灰階影像操取出一腫瘤輪廓及一腫瘤輪廊環形區域，且 此腫瘤輪廓係位於此腫瘤輪廓環形區域内；（B)將此腫瘤輪 廓重疊顯示於此灰階影像上，以在此灰階影像上定義出一 腫瘤内部區域及一腫瘤外部區域；（C)擷取此腫瘤輪廊環形 區域的一重心點，定義一從此重心點向外延伸並通過此腫 瘤輪廊環形區域的剖面線’及提供一位於此剖面線上並位 於此腫瘤輪廓環形區域内之量測線段；（D)計算出位於此量 測線段上之此等像素點所分別具有之灰階移動變異值；以 及（E)依據位於此量測線段上之每一此等像素點所分別具有 之灰階移動變異值，將位於此剖面線上之腫瘤邊緣特徵量 化。 13 201126260 本發明提供—種電腦可讀取記錄媒體，係儲存有用以 使一電腦系統執行一腫瘤邊緣特徵之影像化方法的程式， 此腫瘤邊緣特徵之影像化方法係應用於一由複數個像素點 組合而成並至少顯示一腫瘤的灰階影像，包括下列步驟： (A)從此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環形區 域，且此腫瘤輪廓係位於此腫瘤輪廓環形區域内；（B)將此 腫瘤輪廓重疊顯示於此灰階影像上，以在此灰階影像上定 義出一腫瘤内部區域及一腫瘤外部區域；（c)棟取此腫瘤輪 廓環形區域的一重心點，定義一從此重心點向外延伸並通 過此腫瘤輪廓環形區域的剖面線，及提供一位於此剖面線 上並位於此腫瘤輪廓環形區域内之量測線段；（D)計算出位 於此量測線段上之此等像素點所分別具有之灰階移動變異 值’以及（E)依據位於此量測線段上之此等像素點所分別具 有之灰階移動變異值，定義出一邊緣成像上界及一邊緣成 像下界’且配合一彩虹色階將位於此剖面線上之腫瘤邊緣 特徵影像化。 本發明提供一種電腦可讀取記錄媒體，係儲存有用以 使一電腦系統執行一腫瘤囊腫特徵之量化方法的程式，此 腫瘤囊腫特徵之量化方法係應用於一由複數個像素點組合 而成並至少顯示一腫瘤的灰階影像，包括下列步驟：（A)從 此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環形區域，且 此腫瘤輪廓係位於此腫瘤輪廓環形區域内；將此腫瘤輪 廓重疊顯示於此灰階影像上，以在此灰階影像上定義出一 腫瘤内部區域及一腫瘤外部區域；（C)藉由位於此腫瘤内部 14 201126260 區域内之此等像素點所分別具有之灰階梯度值，計算出位 於此腫瘤内部區域内之此等像素點所具有之灰階梯度值的 最小值及灰階梯度值的標準差；以及（D)依據位於此腫瘤内 部區域内之此等像素點所具有之灰階梯度值的最小值及灰 階梯度值的標準差’將位於此腫瘤内部區域内之囊腫特徵 量化。 本發明提供一種電腦可讀取記錄媒體，係儲存有用以 使一電腦系統執行一腫瘤囊腫特徵之影像化方法的程式， 此腫瘤囊腫特徵之影像化方法係應用於一由複數個像素點 組合而成並至少顯示一腫瘤的灰階影像，包括下列步驟： (A)k此灰階影像操取出一腫瘤輪廓及一腫瘤輪廓環形區 域，且此腫瘤輪廓係位於此腫瘤輪廓環形區域内；（B)將此 腫瘤輪廓重豐顯示於此灰階影像上，以在此灰階影像上定 義出一腫瘤内部區域及一腫瘤外部區域；（c)藉由位於此腫 瘤内部區域内之此等像素點所分別具有之灰階梯度值，計 算出位於此Μ瘤内部區域内之此等像素點所具有之灰階梯 度值的最小值及衣階梯度值的標準差；以及（D)依據位於此 腫瘤内部區域内之此等像素點所具有之灰階梯度值的最小 值及灰階梯度值的標準差，定義出一囊腫成像上界及一囊 腫成像下界，以將位於此腫瘤内部區域内之腫瘤囊腫特徵 影像化。 本發明提供一種電腦可讀取記錄媒體，係儲存有用以 使一電腦系統執行一腫瘤鈣化特徵之量化方法的程式，此 腫瘤鈣化特徵之量化方法係應用於一由複數個像素點組合 201126260 而成並至少顯示一腫瘤的灰階影像，包括下列步驟：（八）從 此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環形區域，且 此腫瘤輪廓係位於此腫瘤輪廓環形區域内；（B)將此腫瘤輪 廓重疊顯示於此灰階影像上，以在此灰階影像上定義出: 腫瘤内部區域及一腫瘤外部區域；（c)藉由位於此腫瘤内部 區域内之此等像素點所分別具有之灰階梯度值，計算出位 於此腫瘤内部區域内之此等像素點所具有之灰階梯度值的 最小值及灰階梯度值的標準差；（D)依據位於此腫瘤内部區 域内之此等像素點所具有之灰階梯度值的最小值及灰階梯 度值的標準差，從此灰階影像中擷取出一位於此腫瘤内部 區域内之囊腫區域；（E)藉由位於此腫瘤内部區域内但位於 此囊腫區域之外之此等像素點所分別具有之灰階梯度值， 計算出位於此腫瘤内部區域内但位於此囊腫區域之外之此 等像素點所具有之灰階梯度值的最大值、灰階梯度值的標 準差及灰階梯度值的平均值；以及（F)依據位於此腫瘤内部 區域内但位於此囊腫區域之外之此等像素點所具有之灰階 梯度值的最大值、灰階梯度值的標準差及灰階梯度值的平 均值’將位於此腫瘤内部區域内之腫瘤鈣化特徵量化。 本發明提供一種電腦可讀取記錄媒體，係儲存有用以 使一電腦系統執行一腫瘤鈣化特徵之影像化方法的程式， 此腫瘤的化特徵之影像化方法係應用於一由複數個像素點 組合而成並至少顯示一腫瘤的灰階影像，包括下列步驟： (A)從此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環形區 域’且此腫瘤輪廓係位於此腫瘤輪廓環形區域内：（B)將此 201126260 腫瘤輪廓重疊顯示於此灰階影像上，以在此灰階影像上定 義出一腫瘤内部區域及一腫瘤外部區域；（c)藉由位於此腫 瘤内部區域内之此等像素點所分別具有之灰階梯度值，計 算出位於此腫瘤内部區域内之此等像素點所具有之灰階梯 度值的最小值及夜階梯度值的標準差；（D)依據位於此腫瘤 内部區域内之此等像素點所具有之灰階梯度值的最小值及 灰階梯度值的標準差，從此灰階影像中擷取出一位於此腫 瘤内部區域内之囊腫區域；（E)藉由位於此腫瘤内部區域内 但位於此囊腫區域之外之此等像素點所分別具有之灰階梯 度值’计异出位於此腫瘤内部區域内但位於此囊腫區域之 外之此等像素點所具有之灰階梯度值的最大值、灰階梯度 值的標準差及灰階梯度值的平均值；以及（F)依據位於此腫 瘤内部區域内但位於此囊腫區域之外之此等像素點所具有 之灰階梯度值的最大值、灰階梯度值的標準差及灰階梯度 值的平均值，定義出一鈣化成像上界及一鈣化成像下界， 以將位於此腫瘤内部區域内之腫瘤鈣化特徵影像化。 本發明提供一種電腦可讀取記錄媒體，係儲存有用以 使一電腦系統執行一腫瘤迴音性特徵之量化方法的程式， 此腫瘤迴音性特徵之量化方法係應用於一由複數個像素點 組合而成並至少顯示一腫瘤的灰階影像，包括下列步驟： (A)從此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環形區 域，且此腫瘤輪廓係位於此腫瘤輪廓環形區域内；（B)將此 腫瘤輪廓重疊顯示於此灰階影像上，以在此灰階影像上定 義出一腫瘤内部區域及一腫瘤外部區域；（c)藉由位於此腫 201126260 瘤内部區域内之此等像素點所分別具有之灰階梯度值，計 算出位於此腫瘤内部區域内之此等像素點所具有之灰階梯 度值的平均值：（D)於此腫瘤外部區域選取一基準區域，藉 由位於此基準區域内之此等像素點所分別具有之灰階梯度 值’計异出位於此基準區域内之此等像素點所具有之灰階 梯度值的平均值；以及（E)依據位於此腫瘤内部區域内之此 等像素點所具有之灰階梯度值的平均值及位於此基準區域 内之此等像素點所具有之灰階梯度值的平均值，將此腫瘤 之迴音性特徵量化。 本發明提供一種電腦可讀取記錄媒體，係儲存有用以 使一電腦系統執行一腫瘤異質化特徵之量化方法的程式， 此腫瘤異質化特徵之量化方法係應用於一由複數個像素點 組合而成並至少顯示一腫瘤的灰階影像，包括下列步驟： (A)從此灰階影像掏丨出一腫瘤輪靡及一腫瘤輪廊環形區 域，且此腫瘤輪廓係位於此腫瘤輪廓環形區域内；將此 腫瘤輪廓重㈣示於此灰階影像上’以在此灰階影像上定 義出-腫瘤内部區域及—腫瘤外部區域；（c)將位於此腫瘤 内部區域内之此等像素點分別定義為複數個參考遮罩且 每-此等參考遮罩係包含—基準像素點與複數個相鄰於此 基準像素點之像素點；⑼計算出每—此等參考遮革所分別 具有之參考遮罩灰階梯度值局部平均及參考遮罩灰階梯度 =部變異；⑹計算出每-此等參考料所分別具有 考遮罩灰階梯度值局部平均之變異、參考遮罩灰階梯度值 局部變異之平均以及參考遮罩灰階梯度值局部變里之變 201126260 異；以及（F)藉由每一此等參考遮罩所分別具有之至少—選 自於一由參考遮箪灰階梯度值局部平均之變異'參考遮= 灰階梯度值局部變異之平均以及參考遮罩灰階梯度值局部 變異之變異所構成之群組，計算出每—此等參考遮罩所分 別具有之異質化指標值，將此腫瘤之異質化特徵量化。 本發明提供-種電腦可讀取記錄媒體，係儲存有用以 使-電腦系統執行-腫瘤異質化特徵之影像化方法的程 式，此腫瘤異質化特徵之影像化方法係應用於一由複數個 像素點組合而成並至少顯示一腫瘤的灰階影像，包括下列 步驟：⑷從此灰階影像操取出一腫瘤輪廊及_腫瘤輪廊環 形區域，且此腫賴廓係、位於此腫瘤輪廓環形區域内；⑻ 將此腫瘤輪廊重憂顯示於此灰階影像上，以在此灰階影像 上定義出-腫瘤内部區域及—腫瘤外部區域；（〇將位於此 腫瘤内部區域内之此等像素點分別定義為複數個參考遮 =且每-此等參考遮罩係包含_基準像素點與複數個相 此基準像素點之像素點；⑼藉由包含於每一此等參考 遮罩之此等像素點所分別具有之灰階梯度值計算出每— * ^考遮罩所77別具有之參考遮罩灰階梯度值變里；（E) ^每二等參考遮罩所分別具有之參考遮罩灰階梯度值 '.计t出此等參考遮罩所具有之平均遮罩灰階梯度值 受”（F)藉由每-此等參考遮罩所分別具有之參考遮罩灰 度值變f及此等參考遮罩所具有之平均遮罩灰階梯度 又異’计异出每_此等參考遮罩所分別具有之異質化指 (G)藉由每—此等參考遮革所分别具有之異質化指標 19 201126260 值，计异出此等參考遮罩所具有之異質化指標值的最大 值、異質化指標值的最小值、異質化指標值的平均值及異 質化指標值的標準差；以及（H)依據此等參考遮罩所具有之 異質化指標值的最大值、異質化指標值的最小值、異質化 指標值的平均值及異質化指標值的標準差，定義出一異質 化成像上界及一異質化成像下界，且配合一彩虹色階將位 於此腫瘤内部區域内之腫瘤異質化特徵影像化。 因此，藉由本發明所提供之腫瘤邊緣特徵的量化方 法、腫瘤邊緣特徵的影像化方法、腫瘤囊腫特徵的量化方 法、腫瘤囊腫特徵的影像化方法、腫瘤鈣化特徵的量化方 法、腫瘤鈣化特徵的影像化方法、腫瘤迴音性特徵的量化 方法、腫瘤異質化特徵的量化方法及腫瘤異質化特徵的影 像化方法，醫師可於拿到一腫瘤之超音波灰階影像的同 時，一併得到腫瘤這些特徵的量化數據與影像化圖像，做 為判斷腫瘤之性質的依據，以大幅提昇藉由腫瘤之超音波 灰階影像判斷腫瘤性質之程序的準確率及可靠度，且減輕 醫師在判斷腫瘤性質時的負擔。 【實施方式】 圖2係顯不一電腦系統之架搆的示意圖，其可用以執行 本發明之腫瘤邊緣特徵的量化方法、腫瘤邊緣特徵的影像 化方法、腫瘤囊腫特徵的量化方法、腫瘤囊腫特徵的影像 化方法、腫瘤鈣化特徵的量化方法、腫瘤鈣化特徵的影像 20 201126260 化方法、腫瘤迴音性特徵的量化方法、腫瘤異質化特徵的 量化方法及腫瘤異質化特徵的影像化方法。 如圖2所示’電腦系統包含顯示裝置2丨、處理器22、記 憶體23、輸入裝置24及儲存裝置25等◊其中，輸入裝置24 可用以輸入影像、文字、指令等資料至電腦系統，儲存裝 置25係例如為硬碟、光碟機或藉由網際網路連接之遠端資 料庫，用以儲存系統程式、應用程式及使用者資料等，記 憶體23係用以暫存資料或執行之程式，處理器22用以運算 及處理資料等，顯示裝置21則用以顯示輸出之資料。如圖2 所示之電腦系統一般係於系統程式26下執行各種應用裎 式，例如文書處理程式、繪圖程式、科學運算程式、瀏覽 程式、電子郵件程式等。 在本實施例中，儲存裝置25係儲存有使電腦系統執行 一腫瘤邊緣特徵之量化方法的程式、使一電腦系統執行— 腫瘤邊緣特徵之影像化方法的程式、使一電腦系統執行一 腫瘤囊腫特徵之量化方法的程式、使一電腦系統執行一腫 瘤囊腫特徵之影像化方法的程式、使一電腦系統執行一腫 瘤妈化特徵之量化方法的程式、使一電腦系統執行一腫瘤 妈化特徵之影像化方法的程式、使一電腦系統執行一腫瘤 迴音性特徵之量化方法的程式 '使一電腦系統執行一腫瘤 異質化特徵之量化方法的程式以及使一電腦系統執行—腫 瘤異質化特徵之影像化方法的程式。當欲使電腦系統執行 某一量化方法或影像化方法時，對應之程式便被載入記情 體23，以配合處理器22執行此量化方法或影像化方法。最 201126260 示裝置21或藉由網 後’再將量化或影像化的結果顯示於顯 際網路儲存於一遠端資料庫中。 歸2外’預備破量化或f彡像化之超音波m像係儲存 、：？裝置25 ’且在被量化或影像化時從储存裝置25被载 入記憶體23，以執行預定之量化方法或影像化方法所包含 的各個步驟。除此之外，腫瘤輪廓擷取方法之「初始腫瘤 輪廓線」係藉由輸入裝置24輸入至電腦系統中，再與超音 波灰階影像互相結合，以執行後續的步驟。 圖3A係一超音波灰階影像的示意圖，其係由複數個像 素點組合而成，且每一像素點分別具有一灰階梯度值。而 如圖3A所示，此超音波灰階影像係顯示一甲狀腺腫瘤與其 周圍的曱狀腺組織。 其次’如圖3B所示，其係本發明第一實施例之腫瘤邊 緣特徵之量化方法的流程圖，其包括下列步驟： (A)從此灰階影像掏取出一腫瘤輪廓及一腫瘤輪廓環 形區域’且此腫瘤輪廓係位於此腫瘤輪廓環形區域内； (B)將此腫瘤輪廓重疊顯示於此灰階影像上，以在此灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C)操取此腫瘤輪廓環形區域的一重心點，定義一從此 重心點向外延伸並通過此腫瘤輪廓環形區域的剖面線，及 提供一位於此剖面線上並位於此腫瘤輪廓環形區域内之量 測線段； (D)計算出位於此量測線段上之每一此等像素點所分 別具有之灰階移動變異值；以及 201126260 (E)依據位於此量測線段上之每一此等像素點所分別 具有之灰階移動變異值，將位於此剖面線上之腫瘤邊緣特 徵量化。 ’、 請參閱圖3C及圖3D，一用於擷取前述步驟（八）之「腫 瘤輪廓」與「腫瘤輪廓環形區域」之「腫瘤輪廓擷取方法」、 步驟（A)所擷取出之「腫瘤輪廓」與「腫瘤輪廓環形區域」 以及步驟（C)之位於「腫瘤内部區域」内之「重心點」與「剖 面線」的定義將敘述於下。其中，圖3C係腫瘤輪廓擷取方 法的流程圖，圖3D則為應用此「腫瘤輪廓擷取方法」以擷 取「腫瘤輪廓」與「腫瘤輪廓環形區域」之包含一腫瘤之 灰階影像圖。 如圖3C所示，在本實施例中，一用於擷取步驟（⑷之「腫 瘤輪廓」與「腫瘤輪廓環形區域」之「腫瘤輪廓擷取方法」 所.應用之腫瘤輪廓擷取方法係包括下列步驟： 輸入一初始腫瘤輪廓線（圖3D中的軌跡3 1); 藉由此初始腫瘤輪廓線定義出一初始腫瘤輪廓環形區 域（由圖3D中的轨跡32與執跡33所包圍的區域），且此初始 腫瘤輪廓係位於此初始腫瘤輪廓環形區域内； 藉由此初始腫瘤輪廓環形區域定義出一初始重心點 (圖3D中的點34)及一從此初始重心點向外延伸並通過此初 始腫瘤輪廓環形區域的初始剖面線（圖3D中的線段35)，及 提供一位於此初始剖面線上並位於此腫瘤輪廓環形區域内 之初始量測線段； 201126260 依據位於此初始量測線段上之此等像素點像所分別具 有之灰階梯度值，計算出位於此初始量測線段上之此等像 素點所分別具有之灰階移動變異值； 比較位於此初始量測線段上之此等像素點所分別具有 之灰階移動變異值，將具有最大之灰階移動變異值的像素 點定義為一位於此初始剖面線上之腫瘤輪廟建議點（圖3 〇 中的點36);以及 改變此初始剖面線之位置以掃瞄此腫瘤之全部邊緣， 且將從不同初始剖面線之位置所分別定義出之複數個腫瘤 輪廓建議點互相連接’以得出此腫瘤輪廓及此腫瘤輪靡環 形區域。 而當從圖3D之灰階影像中傑取出腫瘤輪廓以後，圖3d 之灰階影像中被腫瘤輪廓包圍的部分即為「腫瘤内部區 域」，而圖3D之灰階影像中其他非屬「腫瘤内部區域」的 部分即為「腫瘤外部區域」。 至於如何從位於此初始量測線段上之此等像素點所分 別具有之灰階梯度值計算出位於此初始量測線段上之此等 像素點所分別具有之灰階移動變異值（m〇ving vadance，以下 將以表示）的方法，將配合下列表1敘述於下： 首先，在本實施例中，某一像素點所具之灰階移動變 異值係定義為此像素點所對應之「局部區段」内灰階 梯度^:異值與此像素點所對應之「局部區段」内各「移動 區間」灰階梯度平均變異之比值，即 24 201126260201126260 VI. Description of the Invention: [Technical Field of the Invention] The present invention relates to a method for quantifying tumor features and an imaging method, particularly a feature suitable for the edge features, cyst characteristics, and calcification characteristics of tumors. And quantification methods for quantifying heterogeneous features and imaging methods for imaging these tumor features. [Prior Art] In recent years, medical ultrasound imaging technology has been significantly improved in terms of image resolution and digital image, so medical ultrasound imaging technology is applied to the monitoring of fetal growth. In addition, it is gradually applied to the judgment of various types of tumor states, such as goiter, and due to the non-invasive imaging characteristics of medical ultrasound imaging technology, physicians are gradually assisted by medical ultrasound imaging technology. To determine the neoplastic shellfish and S to assess the subsequent treatment. The first step in judging the nature of the tumor in the ultrasound image of the surgeon's tumor is to identify the contour of the tumor to define the inner region of the tumor and the outer region of the tumor. Then, the physician can distinguish the characteristics of the tumor such as the edge features, cyst characteristics, deuteration characteristics, echogenic characteristics and heterogeneity (4) from the knife corresponding to the super-sound U image in the inner region of the tumor. Quality reference. However, the current medical ultrasound imaging system only allows. The afternoon doctor visually observed what he thought of the tumor contour and then used the handwriting input device to input the tumor to the ultrasound image of the tumor. But 'S is this procedure, there are many unreliable things. 201126260 Because the current method needs to rely entirely on the subjective feelings and experience of the doctor, even the mental state of the doctor at that time, so the tumor contours entered by different physicians for the same-surgical ultrasound image are not the same, as shown in Figure ia. : To, even with the same physician, the contours of the tumors input to the ultrasound images of the same tumor at different times are not the same. Then, the physician can use the assistance of the tumor contour to visually identify whether the inner region of the tumor has various characteristics and the ratio of various features to the inner region of the tumor. Finally, the t teacher again uses the results collected, the distribution of a characteristic of Shikou or a certain characteristic to occupy the entire tumor: the proportion of the region, etc. #1! The nature of the tumor is broken. That is to say, there is no objective mechanism for judging the nature of tumors by ultrasound images of tumors, and the cases of misdiagnosing the nature of tumors are still caused by the technology of judging the nature of tumors by medical ultrasound imaging technology. Can not be trusted by the medical community and the public. In addition, although there are several image edge recognition methods in the field of image recognition (such as license plate recognition), such as the snake algorithm. However, in this case, the algorithm must rely on the user to input a starting edge (ie, the physician still has to manually input a rough outline of the tumor) into the algorithm, and the snake algorithm can begin the subsequent calculation process. Moreover, due to the nature of the 5 stupid algorithms, it is more suitable for applications in the case of obvious image edges, otherwise the results of the calculations often differ greatly from the actual edges. However, the edge of the tumor is generally not obvious, so even if the snake algorithm is applied to the ultrasound image of the tumor, the resulting tumor contour often has a gap with the actual contour of the tumor, as shown in Figure B. Show. Moreover, in order to calculate that the contour of 201126260 is closer to the actual contour of the tumor and shorten the time required for the calculation, the physician also needs to carefully input a starting edge that is not too far from the actual tumor contour, and the result is still not relieved by the physician too much. The burden. In addition, because the ultrasound image of the tumor is a grayscale image, the characteristics of the tumor (such as edge features, cyst characteristics, calcification characteristics 'echo characteristics and heterogeneous characteristics, etc.) are often here. In the grayscale image, only some of the grayscale values of some pixels have slight changes. For the naked eye of the doctor, these features are not easily recognized. The physician can only judge whether these features exist by "feeling". 'The judgment that causes the nature of the tumor can only be based on the subjective feeling of the physician' and cannot be judged accurately based on _. Therefore, the industry needs a method for quantifying tumor features and corresponding imaging methods, especially for quantifying methods for quantifying the features, occupational features, echogenic features and heterogeneous features of marginal features of tumors. An imaging method for imaging tumor features. SUMMARY OF THE INVENTION The main object of the present invention is to provide a method for quantifying tumor characteristics, which can be used to align the tumor with a variety of features, such as edge features, signs, reduction features, The echogenicity is ^ ^ ^ ^ ^ ^ ^ ^ for evaluation. Inch R and heterogeneity characteristics are degenerated for the physician's other purpose of the invention. The 仏 is used in a visualization method for a tumor feature, which means that the tumor has various features of ns H, such as edges. Features, cysts and features of the cysts are visualized for physician evaluation. 6 201126260 In order to achieve the above object, the method for quantifying tumor edge features of the present invention is applied to a gray scale image composed of a plurality of pixel points and displaying at least one tumor, comprising the following steps: (A) from this gray scale image Extracting a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the grayscale image to define the grayscale image An inner region of the tumor and an outer region of the tumor; (C) a center of gravity of the annular region of the contour of the tumor, defining a section line extending outward from the center of gravity and passing through the annular region of the contour of the tumor, and providing a profile a measurement line segment on the line and located in the annular region of the tumor contour; (D) calculating a gray-scale movement variation value of the pixels respectively located on the measurement line segment; and (E) according to the measurement line segment Each of the above pixels has a gray-scale movement variation value, and the tumor edge features located on the section line are quantized. In order to achieve the above object, the imaging method of the tumor edge feature of the present invention is applied to a grayscale image composed of a plurality of pixel points and displaying at least one tumor, including the following steps: (A) from this grayscale image撷Extracting a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the gray image to define the grayscale image An inner region of the tumor and an outer region of the tumor; (C) a center of gravity of the annular region of the tumor wheel, defining a section line extending outward from the center of gravity and passing through the annular region of the contour of the tumor, and providing a a measurement line segment on the section line and located in the annular region of the tumor contour; (D) calculating a gray-scale movement variation value of each of the pixels located on the measurement line segment; and (E) based on the measurement 201126260 The gray-scale movement variation values of these pixels on the line segment respectively define an edge imaging upper bound and an edge imaging lower bound, and match an iridescent color. The characteristic image of the tumor edge line of this cross section. In order to achieve the above object, the method for quantifying the characteristics of the tumor cyst of the present invention is applied to a gray scale image which is composed of a plurality of pixel points and displays at least one tumor, and includes the following steps: (A) taking out the gray scale image a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the gray scale image to define a tumor on the gray scale image An inner region and an outer region of the tumor; (C) calculating, by the gray gradient values of the pixels located in the inner region of the tumor, the gray points of the pixels located in the inner region of the tumor a minimum value of the step value and a standard deviation of the gray step value; and (D) a minimum value of the gray step value and a standard deviation of the gray step value according to the pixels located in the inner region of the tumor, The cyst features located within the inner region of the tumor are quantified. In order to achieve the above object, the 'imaging method of the tumor cyst feature of the present invention' is applied to a gray scale image composed of a plurality of pixel points and displaying at least one tumor, comprising the following steps: (A) from this gray scale image撷Extracting a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) superimposing the tumor contour on the gray scale image to define the gray scale image An inner region of the tumor and an outer region of the tumor; (C) calculating the pixel points located in the inner region of the tumor by the gray gradient value of the pixels located in the inner region of the tumor The minimum value of the gray gradient value and the standard deviation of the gray gradient value 8 201126260; and (D) the minimum value of the gray gradient value and the gray gradient according to the pixels located in the inner region of the tumor The standard deviation of the values defines a cyst imaging upper bound and a cyst imaging lower bound to visualize tumor cyst features located within the tumor's internal region. In order to achieve the above object, the method for quantifying tumor malignant features of the present invention is applied to a grayscale image composed of a plurality of pixels and displaying at least one tumor, including the following steps: (A) from this grayscale image撷Extracting a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) superimposing the tumor contour on the grayscale image to define a grayscale image An internal region of the tumor and an outer region of the tumor; (C) calculating, by the gray gradient values of the pixels located in the interior region of the tumor, the pixels located in the inner region of the tumor The minimum value of the gray gradient value and the & margin of the gray gradient value; (D) the minimum value of the gray gradient value and the gray gradient value according to the pixels located in the inner region of the tumor Poor, a cystic region in the inner region of the tumor is extracted from the grayscale image; (E) by the pixels located in the inner region of the tumor but outside the region of the cyst The gray gradient value respectively is calculated, and the maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the gray ladder of the pixels located in the inner region of the tumor but outside the cyst region are calculated. The mean value of the degree' and (F) the maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the gray ladder according to the pixels located in the inner region of the tumor but outside the cyst region. The mean value of the degrees quantifies the tumor calcification features located within the interior region of the tumor. 201126260 In order to achieve the above object, the imaging method for tumor calcification characteristics of the present invention is applied to a gray scale image composed of a plurality of pixel points and displaying at least one tumor, comprising the following steps: (A) from this gray scale image Extracting a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) superimposing the tumor contour on the grayscale image to define on the grayscale image An inner region of the tumor and an outer region of the tumor; (C) calculating, by the gray gradient values of the pixels located in the inner region of the tumor, the pixels located in the inner region of the tumor have The minimum value of the gray gradient value and the standard deviation of the gray gradient value; (D) the minimum value of the gray gradient value and the standard deviation of the gray gradient value according to the pixels located in the inner region of the tumor , from which a cystic region in the inner region of the tumor is removed from the grayscale image; (E) by being located outside the tumor but outside the cystic region The gray step value of each pixel is calculated, and the maximum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor but outside the cyst region are calculated. And the average value of the gray gradient value; and (F) the maximum value of the gray gradient value and the standard deviation of the gray gradient value according to the pixels located in the inner region of the tumor but outside the cyst region And the average value of the gray gradient values, defining a chemical imaging upper bound and a calcified imaging lower bound 'to visualize the tumor calcification features located in the inner region of the tumor. In order to achieve the above object, the 'quantification method of the tumor echogenic feature of the present invention' is applied to a grayscale image composed of a plurality of pixel points and displaying at least one tumor, including the following steps: (8) From this grayscale image撷Take out a 10 201126260 tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) overlay the tumor contour on the gray scale image to define on the night image An inner region of the tumor and an outer region of the tumor; (C) calculating the pixel points located in the inner region of the tumor by the gray gradient values respectively obtained by the pixels located in the inner region of the tumor Having an average value of the gray gradient value; a reference region is selected in the outer region of the tumor, and the gray gradient value respectively obtained by the pixels located in the reference region is calculated to be located in the reference region The average value of the gray gradient values of the pixels; and (E) the gray ladder of the pixels located in the inner region of the tumor The average of the values and the average of the gray gradient values of the pixels located in the reference region quantify the echogenic features of the tumor. In order to achieve the above, the method for quantifying tumor heterogeneity of the present invention is applied to a gray scale image composed of a plurality of pixel points and displaying at least one tumor, comprising the following steps: (A) grayscale image from this Extracting a tumor contour and a tumor contour annular region ' and the tumor contour is located in the annular contour region of the tumor contour; (B) overlaying the tumor contour on the gray scale image to define on the gray scale image An internal region of the tumor and an external region of the tumor; (C) defining the pixels located within the interior region of the tumor as a plurality of reference masks, and each of the reference masks includes a reference pixel and a plurality of reference pixels a pixel adjacent to the reference pixel; calculating a local average of the reference mask gray gradient value and a local variation of the reference mask gray gradient value for each of the reference masks; (ε) Each - these reference masks have a reference material ash (four) degree value local average 201126260 father's different value of the test mask gray step value local variation and the test mask gray gradient The variation of the local variation of the value; and (F) each of the reference masks having $, respectively, at least one selected from a reference mask gray scale value. The variation of p average, the average of the local variation of the reference mask gray gradient value is the heterogeneity index value of the reference mask composed of the variation of the local variation of the mask gray & gradient value, respectively. The heterogeneous features of this tumor were quantified. For the above (4), the imaging heterozygous feature of the present invention is a gambling cancer smear: for combining a plurality of pixel points and displaying at least a tumor 2 jersey image comprising the following steps: (4) from this grayscale image Exercising and: the annular contour of the tumor contour, and the contour of the tumor is located on the order image, so that the second overlap shows the outer region of the gray tumor; (6) the inner region of the tumor and one is defined as a plurality of reference masks respectively And each of the reference == prime points and a plurality of pixel points adjacent to the reference pixel point 2, 3 respectively, the gray gradient value of each of the pixels of the reference masks, Out—the reference mask gray step value variation; (Ε) calculate the reference step value variation by using the reference mask gray step value variation of each 7:L reference; (F) By the variation of the reference mask gray gradient value of each of the reference masks and the variation of the average mask gray gradient values of the reference masks, the cover is respectively provided with a servant...s... Out of the mother - these reference masks have different meta-values, (G) by 201 126260 The heterogeneity index value respectively 'calculates the maximum value of the heterogeneous index value, the minimum value of the heterogeneity index value, the average value of the heterogeneity index value, and the standard of the heterogeneity index value of the reference masks. And (H) define a maximum value of the heterogeneous index value, the minimum value of the heterogeneity index value, the average value of the heterogeneity index value, and the standard deviation of the heterogeneity index value of the reference masks The heterogeneous imaging upper bound and a heterogeneous imaging lower bound, together with a rainbow gradation, visualize the tumor heterogeneity features located within the interior region of the tumor. The invention provides a computer readable recording medium, which is a program for storing a computer system to perform a tumor edge feature quantization method, wherein the tumor edge feature quantization method is applied to a combination of a plurality of pixel points and Displaying at least a grayscale image of a tumor, comprising the steps of: (A) manipulating a tumor contour and a tumor wheel annular region from the grayscale image, and the tumor contour is located in the annular contour region of the tumor; (B) The contour of the tumor is superimposed on the gray scale image to define an inner region of the tumor and an outer region of the tumor on the grayscale image; (C) extracting a center of gravity of the annular region of the tumor corridor, defining a The center of gravity extends outwardly through the section line ' of the annular region of the tumor wheel and provides a measurement line segment on the line and located in the annular region of the tumor contour; (D) calculates the position on the measurement line segment Gray-scale movement variation values respectively obtained by the pixels; and (E) grays respectively according to each of the pixels located on the measurement line segment Moving variance, this will be the tumor on the section line of the edge feature quantity. 13 201126260 The present invention provides a computer readable recording medium storing a program for causing a computer system to perform a visualization method of a tumor edge feature, the imaging method of the tumor edge feature being applied to a plurality of pixels The points are combined and displayed at least one grayscale image of the tumor, comprising the following steps: (A) extracting a tumor contour and a tumor contour annular region from the grayscale image, and the tumor contour is located in the annular contour region of the tumor; (B) superimposing the contour of the tumor on the grayscale image to define an inner region of the tumor and an outer region of the tumor on the grayscale image; (c) a center of gravity of the annular region of the tumor contour, Defining a section line extending outward from the center of gravity point and passing through the annular region of the tumor contour, and providing a measurement line segment on the section line and located in the annular region of the tumor contour; (D) being calculated on the measurement line segment Each of the pixels has a gray-scale movement variation value 'and (E) respectively according to the pixels located on the measurement line segment There is a gray-scale moving variation value that defines an edge imaging upper bound and an edge imaging lower bound' and images a tumor edge feature on the section line in conjunction with an iridogram. The invention provides a computer readable recording medium, which is a program for storing a computer system to perform a quantitative method for tumor cyst characteristics. The method for quantifying tumor cyst characteristics is applied to a combination of a plurality of pixel points and Displaying at least a grayscale image of a tumor, comprising the steps of: (A) extracting a tumor contour and a tumor contour annular region from the grayscale image, and the tumor contour is located in the annular contour region of the tumor; overlapping the tumor contour Displayed on the grayscale image to define an internal region of the tumor and an external region of the tumor on the grayscale image; (C) each of the pixels located within the region of the tumor 14 201126260 The step value, the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor; and (D) according to the inner region of the tumor The minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixel will be located in the cystic feature in the internal region of the tumor. Of. The invention provides a computer readable recording medium, which is a program for storing a computerized method for performing a tumor cyst feature on a computer system, wherein the imaging method of the tumor cyst feature is applied to a combination of a plurality of pixel points. And displaying at least a grayscale image of a tumor, comprising the steps of: (A) k: the grayscale image manipulates a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; Displaying the contour of the tumor on the grayscale image to define an internal region of the tumor and an external region of the tumor on the grayscale image; (c) by the pixels located in the inner region of the tumor Corresponding to the gray gradient value, the minimum value of the gray gradient value and the standard deviation of the clothing gradient value of the pixels located in the inner region of the tumor; and (D) based on the tumor The minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels in the inner region define a cyst imaging upper bound and a cyst imaging lower bound. The tumor cyst features located in the inner region of the tumor are visualized. The invention provides a computer readable recording medium, which is a program for storing a computer system to perform a quantitative method for tumor calcification. The method for quantifying tumor calcification characteristics is applied to a plurality of pixel point combinations 201126260. And displaying at least a grayscale image of the tumor, comprising the following steps: (8) extracting a tumor contour and a tumor contour annular region from the grayscale image, and the tumor contour is located in the annular contour region of the tumor; (B) The tumor contour is superimposed on the gray scale image to define: the inner region of the tumor and the outer region of the tumor; (c) each of the pixels located in the inner region of the tumor has The gray gradient value, the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor are calculated; (D) according to the inner region located in the tumor The minimum value of the gray gradient value of the pixel and the standard deviation of the gray gradient value, and one of the grayscale images is taken out from the tumor. a cystic region in the region; (E) calculated by the gray gradient value of each of the pixels located in the inner region of the tumor but outside the cyst region, but located in the inner region of the tumor but located in the cyst The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels outside the region; and (F) the cyst located in the inner region of the tumor but located in the cyst The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels outside the region quantify the tumor calcification characteristics located in the inner region of the tumor. The invention provides a computer readable recording medium, which is a program for storing a computerized method for performing a tumor calcification feature on a computer system, and the imaging method of the tumor characteristic is applied to a combination of a plurality of pixel points. And displaying at least one grayscale image of the tumor, comprising the following steps: (A) extracting a tumor contour and a tumor contour annular region from the grayscale image and the tumor contour is located in the annular region of the tumor contour: (B The 201126260 tumor contour is superimposed on the gray scale image to define an internal tumor region and an external tumor region on the grayscale image; (c) by the pixels located in the inner region of the tumor The gray gradient value respectively is calculated, and the minimum value of the gray gradient value and the standard deviation of the night gradient value of the pixels located in the inner region of the tumor are calculated; (D) according to the internal region of the tumor The minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels in the pixel, and one of the grayscale images is taken out from the tumor a cystic region in the region; (E) the gray gradient value of each of the pixels located in the inner region of the tumor but outside the cyst region is located in the inner region of the tumor but is located The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels outside the cyst area; and (F) based on the internal region of the tumor but located The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels outside the cyst region define a calcified imaging upper bound and a calcified imaging lower bound. The calcification features of the tumor located in the inner region of the tumor are visualized. The invention provides a computer readable recording medium, which is a program for storing a computer system to perform a quantification method of tumor echo characteristics, and the method for quantifying tumor echo characteristics is applied to a combination of a plurality of pixel points. And displaying at least a grayscale image of a tumor, comprising the following steps: (A) extracting a tumor contour and a tumor contour annular region from the grayscale image, and the tumor contour is located in the annular contour region of the tumor; (B) The tumor contour is superimposed and displayed on the gray scale image to define an inner region of the tumor and an outer region of the tumor on the gray scale image; (c) the pixels located in the inner region of the tumor in the 201126260 tumor The gray gradient value is respectively calculated, and the average value of the gray gradient values of the pixels located in the inner region of the tumor is calculated: (D) a reference region is selected in the outer region of the tumor, by being located here The gray gradient value of each of the pixels in the reference region is different from the gray ladder of the pixels located in the reference region The average value of the values; and (E) the average value of the gray gradient values of the pixels located in the inner region of the tumor and the gray gradient values of the pixels located in the reference region The mean, quantified the echogenic features of this tumor. The present invention provides a computer readable recording medium storing a program for causing a computer system to perform a quantification method of tumor heterogeneity features applied to a plurality of pixel points combined And displaying at least one grayscale image of the tumor, comprising the following steps: (A) extracting a tumor rim and a tumor rim annular region from the grayscale image, and the tumor contour is located in the annular contour region of the tumor; The tumor contour is weighted (four) on the gray scale image to define the inner region of the tumor and the outer region of the tumor on the gray scale image; (c) the pixels located in the inner region of the tumor are respectively defined a plurality of reference masks and each of the reference masks includes a reference pixel point and a plurality of pixel points adjacent to the reference pixel point; (9) calculating a reference mask for each of the reference masks The local average of the ash gradient value and the reference mask gray gradient = partial variation; (6) Calculate the variation of the local average of each of the reference materials with the mask gray scale value, The average of the local variability of the reference ash gradation value and the local variability of the reference ash gradation value 201126260 are different; and (F) each of the reference masks has at least one selected from the group Calculate each reference frame from the group consisting of the reference occlusion gray step value local average variation 'reference occlusion ash gradation value local variability average and the reference mask gray gradient value local variation variation The heterogeneous index values of the masks respectively quantify the heterogeneous features of the tumor. The invention provides a computer readable recording medium, which is a program for storing a computerized method for performing a tumor heterogeneity feature, and the imaging method of the tumor heterogeneity feature is applied to a plurality of pixels. Point combination and display at least one grayscale image of the tumor, including the following steps: (4) Obtaining a tumor wheel corridor and a tumor ring corridor annular region from the grayscale image, and the swelling contour is located in the annular contour region of the tumor contour (8) display the tumor vertices on this grayscale image to define the inner region of the tumor and the outer region of the tumor on the grayscale image; (the pixels that will be located in the inner region of the tumor) The points are respectively defined as a plurality of reference masks and each of the reference masks includes a _reference pixel point and a plurality of pixels corresponding to the reference pixel point; (9) by being included in each of the reference masks The gray gradient value of each pixel is calculated as the reference mask gray gradient value of each of the 77 masks; (E) ^ each second reference mask has a reference mask cover Step value '. The average mask gray gradient value of the reference masks is determined by (F) the reference mask gray value f and the reference masks respectively provided by each of the reference masks The average mask gray gradient has the same heterogeneity index (G) that each of the reference masks has a heterogeneity index by each of the reference masks. 19 201126260 The value, the maximum value of the heterogeneous index value of the reference mask, the minimum value of the heterogeneity index value, the average value of the heterogeneity index value, and the standard deviation of the heterogeneity index value; and (H) basis The maximum value of the heterogeneous index value, the minimum value of the heterogeneity index value, the average value of the heterogeneity index value, and the standard deviation of the heterogeneity index value of the reference mask define a heterogeneous imaging upper bound and a The heterogeneous imaging lower bound, and with a rainbow gradation, visualizes the tumor heterogeneity features located in the inner region of the tumor. Therefore, the method for quantifying tumor edge features, the imaging method of tumor edge features, and the tumor are provided by the present invention. Cyst Characterization methods, imaging methods for tumor cyst features, methods for quantifying tumor calcification characteristics, imaging methods for tumor calcification features, methods for quantifying tumor echogenic features, methods for quantifying tumor heterogeneity features, and images of tumor heterogeneity features In order to obtain the ultrasound grayscale image of a tumor, the physician can obtain the quantitative data and the image of the tumor as a basis for judging the nature of the tumor, so as to greatly enhance the tumor. The accuracy and reliability of the procedure for judging the nature of the tumor by the ultrasonic grayscale image, and reducing the burden on the physician in judging the nature of the tumor. [Embodiment] FIG. 2 is a schematic diagram showing the architecture of a computer system, which can be used to perform The method for quantifying tumor edge features, the imaging method for tumor edge features, the method for quantifying tumor cyst features, the imaging method for tumor cyst features, the method for quantifying tumor calcification characteristics, and the image of tumor calcification characteristics 20 201126260 Quantification of tumor echogenic features, tumor heterogeneity Method and imaging method for tumor heterogeneity characteristics. As shown in FIG. 2, the computer system includes a display device 2, a processor 22, a memory 23, an input device 24, a storage device 25, etc., and the input device 24 can be used. Input image, text, instructions and other data to the computer system. The storage device 25 is, for example, a hard disk, a CD player or a remote database connected via the Internet for storing system programs, applications and user data. The memory 23 is used for temporarily storing data or executing programs, the processor 22 is used for computing and processing data, and the display device 21 is for displaying output data. The computer system shown in FIG. 2 is generally used in the system program 26 Various application formats are executed, such as a word processing program, a drawing program, a scientific computing program, a browsing program, an email program, etc. In this embodiment, the storage device 25 stores a quantization method for causing the computer system to perform a tumor edge feature. Program that enables a computer system to execute - a method of imaging the edge features of a tumor, enabling a computer system to perform a tumor cyst feature A program for quantifying a method, a program for causing a computer system to perform a visualization method of a tumor cyst feature, a program for causing a computer system to perform a quantitative method of tumor malignant features, and enabling a computer system to perform visualization of a tumor-maturizing feature A program of methods, a program that enables a computer system to perform a method for quantifying a tumor's echogenic characteristics, a program that enables a computer system to perform a method of quantifying a tumor heterogeneous feature, and a computer system to perform an imaging method for tumor heterogeneity features Program. When the computer system is to perform a certain quantization method or imaging method, the corresponding program is loaded into the syllabus 23 to cooperate with the processor 22 to perform the quantization method or the imaging method. Most 201126260 shows the device 21 or displays the result of quantization or visualization on the display network in a remote database. Returning to the 2nd, the ultra-sound m image system of the preparatory break-quantization or f-image is stored, :? The device 25' is loaded from the storage device 25 into the memory 23 when quantized or visualized to perform the various steps involved in the predetermined quantization method or imaging method. In addition, the "initial tumor contour" of the tumor contour extraction method is input to the computer system through the input device 24, and then combined with the ultrasonic grayscale image to perform subsequent steps. Fig. 3A is a schematic diagram of an ultrasonic grayscale image, which is composed of a plurality of pixel points, and each pixel has a gray gradient value. As shown in Fig. 3A, the ultrasonic grayscale image shows a thyroid tumor and the surrounding squamous gland tissue. Next, as shown in FIG. 3B, which is a flowchart of a method for quantifying tumor edge features according to a first embodiment of the present invention, comprising the following steps: (A) extracting a tumor contour and a tumor contour annular region from the grayscale image And the tumor contour is located in the annular region of the tumor contour; (B) the tumor contour is superimposed and displayed on the gray scale image to define an internal tumor region and an external tumor region on the grayscale image; C) taking a center of gravity of the annular region of the tumor contour, defining a section line extending outward from the center of gravity point and passing through the annular region of the tumor contour, and providing a volume on the section line and within the annular region of the tumor contour a line segment; (D) calculating a gray-scale movement variation value for each of the pixels located on the measurement line segment; and 201126260 (E) according to each of the pixels located on the measurement line segment The grayscale movement variation values respectively have quantified the tumor edge features located on the section line. 'Please refer to FIG. 3C and FIG. 3D, one for extracting the "tumor contour" and the "tumor contour extraction method" of the above-mentioned step (8), and the "tumor contour extraction method" and the step (A). The definitions of "central point" and "hatching line" in the "tumor contour" and the "tumor contour area" in step (C) will be described below. 3C is a flow chart of a tumor contour extraction method, and FIG. 3D is a gray scale image image of a tumor including a "tumor contour extraction method" for capturing a "tumor contour" and a "tumor contour annular region". . As shown in Fig. 3C, in the present embodiment, a "tumor contour extraction method" for the extraction step ((4) "tumor contour" and "tumor contour annular region" is used. The applied tumor contour extraction method comprises the steps of: inputting an initial tumor contour (track 3 1 in Fig. 3D); thereby defining an initial tumor contour annular region by the initial tumor contour (from the rail in Fig. 3D) a region surrounded by the trace 32 and the trace 33), and the initial tumor contour is located in the annular region of the initial tumor contour; thereby defining an initial center of gravity point (point 34 in FIG. 3D) An initial section line extending from the initial center of gravity point and passing through the initial annular contour of the tumor contour (line 35 in Fig. 3D), and providing an initial measurement line segment on the initial contour line and located in the annular region of the tumor contour ; 201126260 Calculate the gray-scale movement variation value of each pixel on the initial measurement line segment according to the gray gradient value of the pixel image on the initial measurement line segment; The gray-scale movement variation values of the pixels located on the initial measurement line segment respectively, and the pixel points having the largest gray-scale movement variation value are defined as A tumor round temple suggestion point on this initial section line (point 36 in Figure 3); and changing the position of the initial section line to scan all edges of the tumor, and will be from the positions of the different initial section lines A plurality of defined tumor contours are suggested to be interconnected to obtain the tumor contour and the annular region of the tumor rim. When the tumor contour is taken out from the gray-scale image of Fig. 3D, the portion of the grayscale image of Fig. 3d surrounded by the tumor contour is the "inside of the tumor", and the other grayscale images of Fig. 3D are not tumors. The part of the internal area is the "outside of the tumor". As to how to calculate the gray-scale movement variation values of the pixels located on the initial measurement line segment from the gray gradient values respectively of the pixels located on the initial measurement line segment (m〇ving Vadance, which will be represented by the following method, will be described below with the following list: First, in this embodiment, the gray-scale moving variation value of a certain pixel point is defined as the "partial portion corresponding to the pixel point". The gray level of the segment is ^: the ratio of the difference between the gray value of the "moving interval" in the "local segment" corresponding to the pixel, ie 24 201126260

其中，局部區段係包含此像素點與複數個位於此像素 點之前（或之後）、位於此初始量測線段上的像素點。一般而 言，若P表示在此局部區段中，從此像素點向前或向後包含 之像素點的數目’此局部區段則包含2p+1個像素點。 此外，此局部區段内之移動區間係包含此像素點與複 數個位於此像素點之後、位於此局部區間内之像素點。一 般而言，若q表示在此移動區間中，包含此像素點及位於此 像素點之後之像素點的數目，此局部區段則包含2p_q + 2個 移動區間。 K 1 2 3 4 5 Gijk 77 79 78 79 94 MViJk 4.996216 5.43 1265 5.72553 4.359829 5.065877 k 6 7 8 9 10 Gy'k 93 93 88 78 64 MVijk 8.586214 1 1.91848 13.55545 12.59645 17.43012 k 11 12 13 14 15 Gijk 50 52 56 65 62 MVijk 14.88913 10.87886 6.907269 4.268695 4.830066 k 16 17 18 19 Gijk 60 65 67 69 M^ijk 1 1.60418 13.1099 7.002381 6.10648 201126260 其中，表1中的々表示位於此初始量測線段上之此等像 素點的編號，G供係為此像素點之灰階梯度值，則為此 像素點之灰階移動變異值。 從表1中可看出，編號丨〇的像素點具有最大的灰階移動 變異值，而此位於此初始量測線段上之像素點便為前述之 「腫瘤輪廓建議點」。 此外，如圖3 E所示，在本實施例之腫瘤邊緣特徵的量 化方法中，步驟（〇)包括一步驟（D 1),依據位於此量測線段 上之此等像素點所分別具有之灰階移動變異值，計算 出位於此量測線段上之此等像素點所具有之灰階移動變異 值的標準差（wMK)。至於從灰階移動變異值計算出灰階移動 變異值之標準差的方法，由於已廣為各界所熟悉，在此便 不再贅述。 除此之外，如圖3E所示，步驟（D)於前述之步驟（D1) 之後更包括一步驟（D2) ’依據位於此量測線段上之此等像 素點所分別具有之灰階移動變異值，計算出位於此量 測線段上之此等像素點所具有之灰階移動變異值的平均值 (@)。至於從灰階移動變異值計算出灰階移動變異值之平 均值的方法’由於已廣為各界所熟悉，在此亦不再贅述。 另一方面’如圖3F所示，在本實施例之腫瘤邊緣特徵 的影像化方法中，步驟（E)包括一步驟（E 1)’藉由位於此量 測線段上之此等像素點所具有之灰階移動變異值的標準差 及灰階移動變異值的平均值（77^ )，定義出一灰階移動 26 201126260 異值閥值（threshold value) ’以判別位於此剖面線上之腫瘤邊 緣特徵的模糊程度。 在本實施例中’此閥值係為位於此量測線段上之此等 像素點所具有之灰階移動變異值的的平均值（死)加上三倍 之此灰階移動變異值的標準差，即丽^+3 χ 。若一位於 此量測線段上之像素點所具有之灰階移動變異值係低 於此閥值時，此像素點便被定義為具有邊緣模糊特徵。在 比對完所有位於此量測線段上的像素點後，將被定義為具 有邊緣模糊特徵之像素點的數目除以所有位於此量測線段 上之像素點的數目後，便可得出邊緣模糊特徵於此剖面線 之位於此腫瘤輪廓環形區域内之部分所佔的比例。 最後，本發明第一實施例之腫瘤邊緣特徵之量化方法 可於步驟（E)之後更包括一步驟（F)改變此刳面線之位置，以 掃瞄此腫瘤之全部邊緣而將此腫瘤之全部邊緣特徵量化。 如此，便可得出在腫瘤之全部邊緣中，具有邊緣模糊 之邊緣所佔的比例。 圖4 A係本發明第二實施例之腫瘤邊緣特徵之影像化 方法的流程圖’其包括下列步驟： (A) 從此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環 升> 區域，且此腫瘤輪廓係位於此腫瘤輪廓環形區域内； (B) 將此腫瘤輪廓重疊顯示於此灰階影像上以在此灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 擷取此腫瘤輪廓環形區域的一重心點，定義一從此 重心點向外延伸並通過此腫瘤輪廓環形區域的剖面線，及 27 201126260 提供一位於此剖面線上並位於此腫瘤輪廓環形區域内之量 測線段； (D) 計算出位於此量測線段上之此等像素點所分別具 有之灰階移動變異值；以及 (E) 依據位於此量測線段上之此等像素點所分別具有 之灰階移動變異值，定義出一邊緣成像上界及一邊緣成像 下界，且配合一彩虹色階將位於此剖面線上之腫瘤邊緣特 徵影像化。 其中，由於步驟（A)所擷取出之「腫瘤輪廓」與「腫瘤 輪廓環形區域」、步驟（B)之藉由「腫瘤輪廓」所定義出之 「腫瘤内部區域」與「腫瘤外部區域」、步驟（c)之位於「腫 瘤内部區域」内之「重心點」與「剖面線」以及步驟（D)之 位於「剖面線」上之各像素點所分別具有之灰階移動變異 值的計算方式均已詳細敘述於前，在此便不再重複敘述。 此外’如圖4B所示，在本實施例之腫瘤邊緣特徵的影 像化方法中，步驟（D)包括一步驟（D1)，依據位於此量測線 段上之此等像素點所分別具有之灰階移動變異值（I.)，計 算出位於此量測線段上之此等像素點所具有之灰階移動變 異值的標準差Ο,ΜΚ,.)。至於從灰階移動變異值計算出灰階移 動變異值之標準差的方法，由於已廣為各界所熟悉，在此 便不再贅述。 除此之外’如圖4Β所示，前述之步驟（D1)之後更包括 一步驟（D2) ’依據位於此量測線段上之此等像素點所分別 具有之灰階移動變異值（ΜΚ,)，計算出位於此量測線段上之 28 201126260 此等像素點所具有之灰階移動變異值的平均值（X7j7)。至於 從灰階移動變異值計算出灰階移動變異值之平均值的方 法’由於已廣為各界所熟悉，在此亦不再贅述。 另一方面，如圖4C所示，在本實施例之腫瘤邊緣特徵 的影像化方法中，步驟（E)包括一步驟（E1)，藉由位於此量 測線段上之此等像素點所具有之灰階移動變異值的標準差 及灰階移動變異值的平均值（研丨）定義出此邊緣成像 上界及此邊緣成像下界。 而在本實施例中，「邊緣成像上界」係為此灰階移動 變異值的平均值加上三倍之此灰階移動變異值的標準差 (+3 ，「邊緣成像下界」則為此灰階移動變異值的The local segment includes the pixel and a plurality of pixels located before (or after) the pixel on the initial measurement line segment. In general, if P denotes the number of pixels included in the local segment from the pixel point forward or backward', the local segment contains 2p+1 pixels. In addition, the movement interval in the local segment includes the pixel point and a plurality of pixel points located in the local interval after the pixel point. In general, if q represents the number of pixels in the moving interval that contain the pixel and the pixel after the pixel, the local segment contains 2p_q + 2 moving intervals. K 1 2 3 4 5 Gijk 77 79 78 79 94 MViJk 4.996216 5.43 1265 5.72553 4.359829 5.065877 k 6 7 8 9 10 Gy'k 93 93 88 78 64 MVijk 8.586214 1 1.91848 13.55545 12.59645 17.43012 k 11 12 13 14 15 Gijk 50 52 56 65 62 MVijk 14.88913 10.87886 6.907269 4.268695 4.830066 k 16 17 18 19 Gijk 60 65 67 69 M^ijk 1 1.60418 13.1099 7.002381 6.10648 201126260 where 々 in Table 1 indicates the number of such pixels located on this initial measurement line segment, G is the gray gradient value of the pixel, and the grayscale movement variation value of the pixel is the pixel. As can be seen from Table 1, the pixel of the number 丨〇 has the largest gray-scale movement variation value, and the pixel point on the initial measurement line segment is the aforementioned "tumor contour suggestion point". In addition, as shown in FIG. 3E, in the method for quantifying the edge feature of the tumor in the embodiment, the step (〇) includes a step (D1), respectively, according to the pixels located on the measurement line segment. The gray scale movement variation value is calculated as the standard deviation (wMK) of the gray scale movement variation values of the pixels located on the measurement line segment. As for the method of calculating the standard deviation of the gray-scale moving variation value from the gray-scale moving variation value, since it is widely known, it will not be described here. In addition, as shown in FIG. 3E, step (D) further includes a step (D2) after the step (D1) described above. 'The gray scale movement respectively according to the pixels located on the measurement line segment respectively. The variation value calculates the average value (@) of the gray-scale movement variation values of the pixels located on the measurement line segment. As for the method of calculating the average value of the gray-scale moving variation value from the gray-scale moving variation value, it has been widely known and will not be described here. On the other hand, as shown in FIG. 3F, in the imaging method of the tumor edge feature of the embodiment, the step (E) includes a step (E1)' by the pixels located on the measurement line segment. The standard deviation of the gray-scale moving variation value and the average value of the gray-scale moving variation value (77^), define a gray-scale movement 26 201126260 threshold value (threshold value) to discriminate the tumor edge located on the section line The degree of blurring of features. In the present embodiment, 'this threshold value is the average value (dead) of the gray-scale movement variation values of the pixels on the measurement line segment plus three times the standard of the gray-scale movement variation value. Poor, that is, Li ^+3 χ. A pixel is defined as having an edge blur feature if the grayscale shift variability of a pixel located on the measurement line is below this threshold. After comparing all the pixels located on the measurement line segment, the number of pixels defined as having edge blur features is divided by the number of pixels located on the measurement line segment to obtain the edge. The proportion of the portion of the hatched line that lies within the annular region of the tumor contour. Finally, the method for quantifying the edge feature of the tumor according to the first embodiment of the present invention may further comprise a step (F) after step (E) to change the position of the face line to scan all the edges of the tumor to treat the tumor. Quantify all edge features. In this way, the proportion of edges with blurred edges in all edges of the tumor can be derived. 4A is a flow chart of a method for imaging a tumor edge feature according to a second embodiment of the present invention, which includes the following steps: (A) extracting a tumor contour and a tumor contour ring elevation region from the grayscale image, and The tumor contour is located in the annular region of the tumor contour; (B) the tumor contour is superimposed and displayed on the gray scale image to define an internal tumor region and an external tumor region on the gray scale image; (C) 撷Taking a center of gravity of the annular region of the tumor contour, defining a section line extending outward from the center of gravity point and passing through the annular region of the tumor contour, and 27 201126260 providing a measurement on the contour line and located in the annular region of the tumor contour a line segment; (D) calculating a gray-scale movement variation value of each of the pixels located on the measurement line segment; and (E) a gray scale having the pixels respectively located on the measurement line segment The motion variation value defines an edge imaging upper bound and an edge imaging lower bound, and images a tumor edge feature located on the section line in conjunction with an iridogram. Among them, the "tumor contour" and the "tumor contour annular region" taken out in the step (A), and the "tumor inner region" and the "tumor outer region" defined by the "tumor contour" in the step (B), The calculation method of the gray-scale movement variation value of the "center of gravity" and "hatching" in the "inside of the tumor" and the pixel points on the "hatch line" in the step (D) All have been described in detail before, and the description will not be repeated here. In addition, as shown in FIG. 4B, in the imaging method of the tumor edge feature of the embodiment, the step (D) includes a step (D1), respectively, according to the pixels located on the measurement line segment. The step shift value (I.) calculates the standard deviation 灰, ΜΚ, .) of the gray scale movement variation values of the pixels located on the measurement line segment. As for the method of calculating the standard deviation of the gray scale moving variation value from the gray scale moving variation value, since it is widely known, it will not be described here. In addition, as shown in FIG. 4A, the foregoing step (D1) further includes a step (D2) 'based on the gray-scale movement variation values respectively corresponding to the pixels located on the measurement line segment (ΜΚ, ), calculate the average value (X7j7) of the gray-scale movement variation values of the pixels on the measurement line segment. As for the method of calculating the average value of the gray-scale moving variation value from the gray-scale moving variation value, it has been widely known and will not be described here. On the other hand, as shown in FIG. 4C, in the imaging method of the tumor edge feature of the embodiment, the step (E) includes a step (E1), wherein the pixels located on the measurement line segment have The standard deviation of the gray scale moving variation value and the average value of the gray scale moving variation value (Diller) define the upper boundary of the edge imaging and the lower bound of the edge imaging. In the present embodiment, the "edge imaging upper bound" is the standard deviation of the gray-scale moving variation value plus the three-fold standard deviation of the gray-scale moving variation value (+3, "edge imaging lower bound" is Gray scale moving variation

平均值減去三倍之此灰階移動變異值的標準差X )。但是，若「邊緣成像下界」的數值低於此等像素點 所具有之灰階移動變異值的最小值，則改以此等像素點所 具有之灰階移動變異值的最小值做為「邊緣成像下界」的 數值。 耑注意的是，在其他的應用狀況下，前述之「邊緣成 像上界」及「邊緣成像下界」亦可具有不同的數值，如「邊 緣成像上界」可為灰階移動變異值的平均值加上兩倍之此 灰階移動變異ii的標準差硕+2、规），「邊緣成像下界」 則可為此灰階移動變異值的平均值減去兩倍之此灰階移動 變異值的標準差规），只要「邊緣成像上界」的數 值大於「邊緣成像下界」的數值即可。The mean is subtracted by three times the standard deviation of this gray-scale movement variation value X). However, if the value of the "edge of the edge imaging" is lower than the minimum value of the gray-scale movement variation value of the pixels, the minimum value of the gray-scale movement variation value of the pixels is changed as the "edge". The value of the lower bound of imaging. It should be noted that, under other application conditions, the above-mentioned "edge imaging upper bound" and "edge imaging lower bound" may also have different values, such as "edge imaging upper bound" may be the average value of grayscale moving variation values. Adding twice the standard deviation of the gray-scale moving variation ii +2, the "edge imaging lower bound" can subtract twice the gray-scale moving variation value for the average of the gray-scale moving variation values. Standard deviation gauge), as long as the value of "edge imaging upper bound" is larger than the value of "edge imaging lower bound".

2Q 201126260 此外’在本實施例之腫瘤邊緣特徵的影像化方法中， 步驟（E)所配合的此彩虹色階係為一紅检黃綠藍鼓紫之連續 漸變色階。而在腫瘤邊緣特徵影像化時，每一位於此量測 線段上之像素點係分別依據下述之影像化規則而被影像 化： 1 ·若此位於此量測線段上之像素點所具有之灰階移 動變異值大於或等於前述之「邊緣成像上界」， 則以一紅色區塊覆蓋此像素點； 2. 若此位於此量測線段上之像素點所具有之灰階移 動變異值小於或等於前述之「邊緣成像下界」， 則以一紫色區塊覆蓋此像素點；以及 3. 若此位於此量測線段上之像素點所具有之灰階移 動變異值介於前述之「邊緣成像上界」與前述之 「邊緣成像下界」之間，便依據此像素點所具有 之灰階移動變異值分別與前述之「邊緣成像上界」 與前述之「邊緣成像下界」之間的對應關係，以 一具有從此彩虹色階中對應出之顏色的區塊覆蓋 此像素點。 而在依據上述影像化規則將位於此量測線段上之此等 像素點影像化之後’即完成步驟（E)之後，本實施例之腫瘤 邊緣特徵的影像化方法可更包括一步驟（F)改變此剖面線之 位置’以掃猫此腫瘤之全部邊緣。如此，可將位於此腫瘤 的全部邊緣上之腫瘤輪廓環形區域内的所有像素點影像 化，得出如圖5所示之腫瘤邊緣特徵影像化圖像。而藉由如 30 201126260 圖5所示之腫瘤邊緣特徵影像化圓像醫師可輕易判斷出此 腫瘤之邊緣特徵的分佈及腫瘤邊緣模糊的程度。 …需注意的是’在腫瘤邊緣特徵影像化時，了持續地 將則述之配合衫虹色階而得出之各個顏色區塊覆蓋於此灰 P&〜像以外亦可間歇地顯示這些顏色區塊於此灰階影像 上，以利醫生同時地觀察此腫瘤所具有的其他特徵。 圖6A係一超音波灰階影像的示意圖，其係由複數個像 素點組合而成，|每一像素點分別具有一灰階梯度值。如 圖6A所示，其顯示一甲狀腺腫瘤與其周圍的甲狀腺組織， 且此甲狀腺腫瘤包含一囊腫區域。 圖6B係本發明第三實施例之腫瘤囊腫特徵之量化方法 的流程圖’其包括下列步驟： (A) 從此灰階影像操取出一腫瘤輪廓及一腫瘤輪廓環 形區域，且此腫瘸輪廓係位於此腫瘤輪廓環形區域内； (B) 將此腫瘤輪廓重疊顯示於此灰階影像上，以在此灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 藉由位於此腫瘤内部區域内之此等像素點所分別 具有之灰階梯度值，計算出位於此腫瘤内部區域内之此等 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差；以及 (D) 依據位於此腫瘤内部區域内之此等像素點所具有 之灰階梯度值的最小值及灰階梯度值的標準差，將位於此 腫瘤内部區域内之囊腫特徵1量化。 201126260 其中，由於步驟（A)所擷取出之「腫瘤輪廓」與「腫瘤 輪廓裱形區域」以及步驟（B)之藉由「腫瘤輪廓」所定義出 之「腫瘤内部區域」與「腫瘤外部區域」均已詳細敘述於 月在此便不再重複敘述。 此外，由於步驟（C)之計算出位於「腫瘤内部區域」内 之此等像素點所具有之灰階梯度值的最小值及灰階梯度值 的標準差的方法已廣為各界所熟悉，在此便不再贅述。 除此之外，如圖6C所示，在本實施例之腫瘤囊腫特徵 的量化方法中，步驟（D)包括一步驟（D1)，藉由位於此腫瘤 内部區域内之此等像素點所具有之灰階梯度值的最小 值及灰階梯度值的標準差，定義出一囊腫特徵之 灰階梯度值的閥值（threshold value)，以計算出此囊腫特徵於此 腫瘤内部區域内所佔的比例。 在本實施例中，此閥值係為位於此腫瘤内部區域内之 此等像素點所具有之灰階梯度值（GV/)的最小值（心^/)加上零 點一倍之此灰階梯度值的標準差’即丨χ 。 若一位於此腫瘤内部區域内之此等像素點所具有之灰階梯 度值（G〃/)係低於此閥值時’此像素點便被定義為具有囊腫特 徵。在比對完所有位於此腫瘤内部區域内的像素點後，將 被定義為具有囊腫特徵之像素點的數目除以所有位於此腫 瘤内部區域内之像素點的數目後’便可得出囊腫特徵於此 腫瘤内部區域内所佔的比例。 圖7A係本發明第四實施例之腫瘤囊腫特徵之影像化 方法的流程圖，其包括下列步驟： 32 201126260 (A) 從此灰階影像操取出一腫瘤輪廓及一腫瘤輪廊環 形區域，且此腫瘤輪廓係位於此腫瘤輪廊環形區域内； (B) 將此腫瘤輪廓重疊顯示於此灰階影像上，以在此灰^ 階影像上定義出一腫瘤内部區域及一腫瘤外部區域： (C) 藉由位於此腫瘤内部區域内之此等像素點所分別 具有之灰階梯度值，計算出位於此腫瘤内部區域内之此等 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差；以及 (D) 依據位於此腫瘤内部區域内之此等像素點所具有 之灰階梯度值的最小值及灰階梯度值的標準差，定義出一 囊腫成像上界及一囊腫成像下界，以將位於此腫瘤内部區 域内之腫瘤囊腫特徵影像化。 其中’由於步驟（A)所擷取出之「腫瘤輪廓」與「腫瘤 輪廓環升> 區域」以及步驟（B)之藉由「腫瘤輪廓」所定義出 之「腫瘤内部區域」與「腫瘤外部區域」均已詳細敘述於 前’在此便不再重複敘述。 此外’由於步驟（C)之計算出位於「腫瘤内部區域」内 之此等像素點所具有之灰階梯度值的最小值及灰階梯度值 的標準差的方法已廣為各界所熟悉，在此便不再贅述。 除此之外，在本實施例之腫瘤囊腫特徵的影像化方法 _1_ ’步驟（D)之囊臛成像上界係為位於此腫瘤内部 區域内之 此等像素點所具有之灰階梯度值的最小值（w,.„Gw)加上零 點—倍之此灰階梯度值的標準差（4)，即4/+0.丨X.4。 201126260 另一方面，囊腫成像下界則為位於此腫瘤内部區域内之此 等像素點所具有之灰階梯度值（^)的最小值。 需注意的是，在其他的應用狀況下，前述之「囊腫成 像上界」及「囊腫成像下界」亦可具有不同的數值，如「囊 腫成像上界」可為位於此腫瘤内部區域内之此等像素點所 具有之灰階梯度值的最小值加上零點三倍之此灰階梯度值 的標準差，（即m"7(^+0.3Xi,〆^)，「囊腫成像下界」則可為位 於此腫瘤内部區域内之此等像素點所具有之灰階梯度值的 最小值加上零點零五倍之此灰階梯度值的標準差，（即 ，只要「囊腫成像上界」的數值大於「囊腫 成像下界J的數值即可。 此外，當腫瘤囊腫特徵影像化時，若此位於此腫瘤内 部區域内之此等像素點所具有之灰階梯度值係介於前述之 「囊腫成像上界」與「囊腫成像下界」之間，此像素點便 被桃紅色區塊覆蓋。如此，在將每一位於此腫瘤内部區 域内之此等像素點依據此規則影像化之後，便可得出如圖 7B所示之腫瘤囊腫特徵影像化圖像。而藉由如圖7B所示之 腫瘤囊腫特徵影像化圖像，醫師可輕易判斷出此腫瘤之囊 腫特徵於腫瘤内的分佈及佔腫瘤内部區域的比率。 除此之外’本發明第四實施例之腫瘤囊腫特徵之影像 化方法於步驟（D)後更包括下列步驟： (E)將位於該腫瘤内部區域内之該等像素點分別定義 為複數個參考遮罩，且每一該等參考遮罩係包含一基準像 素點與複數個相鄰於該基準像素點之像素點’且該基準像 34 201126260 素點所具有之灰階梯度值係 成像下界之間；以及 介於該囊腫成像上界及該囊腫 (少）當至少 軎絲；、 “寺像素點所具有之灰階梯度值介於該 ，等像= :腫成像下界之間時，該基準像素點及 =寻像素时腫㈣雜徵料料便被—桃紅色區塊覆 JOL· 尽七明第四實施例之腫瘤囊腫特徵之影 化方法於前述之步驟（F)後更包括—步驟（g)，當只有該基準 像素點所具有之灰階梯度值介於該囊腫成像上界及，玄囊腫 成像下界之間時’移除覆蓋於該基準像素點及該等像素點 的該桃紅色區塊。 μ 如此，藉由完成前述之步驟（Ε)至步驟（G)，桃紅色區 塊的形狀及面積便與腫瘤特徵之實際形狀及實際面積更佳 符合。另外，在本實施例中，步驟（E)中所定義之參考遮罩 係包含9個像素點。 圖8A係一超音波灰階影像的示意圖，其係由複數個像 素點組合而成，且每一像素點分別具有一灰階梯度值。如 圖8A所示，其顯示一曱狀腺腫瘤與其周圍的曱狀腺組織， 且此甲狀腺腫瘤包含一 I弓化區域。 圖8B係本發明第五實施例之腫瘤妈化特徵之量化方法 的流程圖’其包括下列步驟： (A)從此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環 形區域’且此腫瘤輪廓係位於此腫瘤輪廓環形區域内； 201126260 (B) 將此腫瘤輪廓重疊顯示於此灰階影像上，以在此灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 藉由位於此腫瘤内部區域内之此等像素點所分別 具有之灰階梯度值，計算出位於此腫瘤内部區域内之此等 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差； (D) 依據位於此腫瘤内部區域内之此等像素點所具有 之灰階梯度值的最小值及灰階梯度值的標準差，從此灰階 影像中擷取出一位於此腫瘤内部區域内之囊腫區域； (E) 藉由位於此腫瘤内部區域内但位於此囊腫區域之 外之此等像素點所分別具有之灰階梯度值，計算出位於此 腫瘤内部區域内但位於此囊腫區域之外之此等像素點所具 有之灰階梯度值的最大值、灰階梯度值的標準差及灰階梯 度值的平均值；以及 (F) 依據位於此腫瘤内部區域内但位於此囊腫區域之 外之此等像素點所具有之灰階梯度值的最大值、灰階梯度 值的標準差及灰階梯度值的平均值，將位於此腫瘤内部區 域内之腫瘤鈣化特徵量化。 其中’由於步驟（A)所擷取出之「腫瘤輪廓」與「腫瘤 輪廓環形區域」、步驟（B)之藉由「腫瘤輪廓」所定義出之 「腫瘤内部區域」與「腫瘤外部區域」以及步驟（D)所擷取 之「囊腫區域」均已詳細敘述於前，在此便不再重複敘述。 此外’由於步驟（C)之計算出位於「腫瘤内部區域」内 之此等像素點所具有之灰階梯度值的最小值及灰階梯度值 201126260 的標準差的方法、步驟（E)之計算出位於「腫瘤内部區域」 内但位於「囊腫區域」外之此等像素點所具有之灰階梯度 值的最大值、灰階梯度值的標準差及灰階梯度值的平均值 的方法已廣為各界所熟悉’在此便不再贅述。 除此之外，如圖8C所示’在本實施例之腫瘤鈣化特徵 的量化方法中’步驟（F)包括一步驟（F1)，藉由位於此腫瘤 内部區域内但位於此囊腫區域之外之此等像素點所具有之 灰階梯度值（G)/)的最大值'灰階梯度值的標準差 及灰階梯度值的平均值（we(w，定義出一鈣化特徵 之灰階梯度值的閥值(thresho丨d value)，以計算出此鈣化特徵於 此腫瘤内部區域内所佔的比例。 在本貫施例中’此閥值係為位於此腫瘤内部區域内但 位於此囊腫區域之外之此等像素點所具有之灰階梯度值（g、) 的平均值（加上一點八倍之此灰階梯度值的標準差 («<_而/)，即咖—馬+2·8Χ_(7〆若一此腫瘤内部區域内但位 於此囊腫區域之外之此等像素點所具有之灰階梯度值 係高於此閥值時，此像素點便被定義為具有鈣化特徵。在 比對完所有位於此腫瘤内部區域内但位於此囊腫區域之外 之此等像素點後，將被定義為具有鈣化特徵之像素點的數 目除以所有位於此腫瘤内部區域内之像素點的數目後，便 可得出鈣化特徵於此腫瘤内部區域内所佔的比例。 圖9A係本發明第六實施例之腫瘤鈣化特徵之影像化 方法的流程圖，其包括下列步驟： 201126260 (A) 從此灰階影像擷取出一腫瘤輪廓及一腫瘤輪_玉裝 形區域，且此腫瘤輪廓係位於此腫瘤輪廓環形區域内； (B) 將此腫瘤輪廓重疊顯示於此灰階影像上，以在此灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 藉由位於此腫瘤内部區域内之此等像素點所分別 具有之灰階梯度值，計算出位於此腫瘤内部區域内之此等 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差； (D) 依據位於此腫瘤内部區域内之此等像素點所具有 之灰階梯度值的最小值及灰階梯度值的標準差，從此灰階 影像中擷取出一位於此腫瘤内部區域内之囊腫區域； (E) 藉由位於此腫瘤内部區域内但位於此囊腫區域之 外之此等像素點所分別具有之灰階梯度值，計算出位於此 腫瘤内部區域内但位於此囊腫區域之外之此等像素點所具 有之灰階梯度值的最大值、灰階梯度值的標準差及灰階梯 度值的平均值；以及 (F) 依據位於此腫瘤内部區域内但位於此囊腫區域之 外之此等像素點所具有之灰階梯度值的最大值、灰階梯度 值的標準差及灰階梯度值的平均值，定義出一鈣化成像上 界及—鈣化成像下界’以將位於此腫瘤内部區域内之腫瘤 鈣化特徵影像化。 其中，由於步驟（A)所擷取出之「腫瘤輪廓」與「腫瘤 輪廓環形區域」、步驟（B)之藉由「腫瘤輪廓」所定義出之 38 201126260 腫瘤内部區域」與「腫瘤外部區域」以及步驟（D)所摘取 之「囊腫區域」均已詳細欽述於前，在此便不再重複钦述。 此外，由於步驟（c)之計算出位於「腫瘤内部區域」.内 之此等像素點所真有之灰階梯度值的最小值及灰階梯度值 的私準差的方法、步驟（E)之計算出位於「腫瘤内部區域』 内但位於「囊腫區域」外之此等像素點所具有之灰階梯度 值的最大值、灰階梯度值的標準差及灰階梯度值的平均值 的方法已廣為各界所熟悉’在此便不再贅述。 除此之外，在本實施例之腫瘤鈣化特徵的影像化方法 中，步驟（F)之鈣化成像上界係為位於此腫瘤内部區域内但 位於此囊腫區域之外之此等像素點所具有之灰階梯度值的 最大值，巧化成像下界則為位於此腫瘤内部區域内 但位於此囊腫區域之外之此等像素點所具有之灰階梯度值 的平均值加上二點八倍之此灰階梯度值的標準差 ’ 即順^〔(7,)/+2.8 。 需注意的是，在其他的應用狀況下，前述之「鈣化成 像上界」及「鈣化成像下界」亦可具有不同的數值，如「鈣 化成像上界」可為位於此腫瘤内部區域内但位於此囊腫區 域之外之此等像素點所具有之灰階梯度值的最大值 減去零點一倍之此灰階梯度值的標準差，即_〇」 ’ 「的化成像下界」則可為位於此腫瘤内部區域内 但位於此囊腫區域之外之此等像素點所具有之灰階梯度值 的平均值（mra加上二點五倍之此灰階梯度值的標準差 201126260 (i./i/一cO(y/)，即 mean c +2.5 Xs，心.C^y/ ’只要「約化成像上界」的數 值大於「鈣化成像下界」的數值即可。 此外，當腫瘤鈣化特徵影像化時，若此位於此腫瘤内 部區域内但位於此囊腫區域之外之此等像素點所具有之灰 階梯度值介於前述之「鈣化成像上界」與「鈣化成像下界 之間時，此像素點便被一黃色區塊覆蓋。如此，在將每一 位於此腫瘤内部區域内但位於此囊腫區域之外之此等像素 點依據此規則影像化之後，便可得出如圖9B所示之腫瘤|弓 化特徵影像化圖像。而藉由如圖9 B所示之腫瘤約化特徵影 像化圖像，醫師可輕易判斷出此腫瘤之赶化特徵於腫瘤内 的分佈及佔腫瘤内部區域的比率。 圖10A係一超音波灰階影像的示意圖，其係由複數個 像素點組合而成’且每一像素點分別具有一灰階梯度值。 如圖10A所示，其顯示一甲狀腺腫瘤與其周圍的甲狀腺組 織。 圖1 0B係本發明第七實施例之腫瘤迴音性特徵之量化 方法的流程圖，其包括下列步驟： (A) 從此灰階影像操取出一腫瘤輪廓及一腫瘤輪廓環 形區域，且此腫瘤輪廓係位於此腫瘤輪廓環形區域内； (B) 將此腫瘤輪廓重疊顯示於此灰階影像上，以在此灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 藉由位於此腫瘤内部區域内之此等像素點所分別 具有之灰階梯度值，計算出位於此腫瘤内部區域内之此等 像素點所具有之灰階梯度值的平均值； 40 201126260 ⑼於此腫瘤外部區域選#_基準區域，藉由位於此基 準區域内之此等像素點所分別具有之灰階梯度值，吁管出 位於此基準區域内之此等像素點所具有之灰階梯度值：平 均值；以及 ⑹依據位於此腫瘤内部區域内之此等像素點所具有 之灰階梯度值的平均值及位於此基準區域内之此等像素點 所具有之灰階梯度值的平均值，將此義之迴音性特徵量 化。 其中’由於步驟⑷所操取出之「腫瘤輪廓」與「腫瘤 輪廓環形區域」以及步驟（B)之藉由「腫瘤㈣」所定義出 之「腫瘤内部區域」與「腫瘤外部區域」肖已詳細叙述於 前，在此便不再重複敘述。 此外，由於步驟（C)之計算出位於「腫瘤内部區域」内 之此等像素點所具有之灰階梯度值之平均值的方法以及步 驟（D)之计算出位於.基準區域」内之此等像素點所具有之 灰階梯度值之平均值的方法已廣為各界所熟悉，在此便不 再贅述。 除此之外，前述之「基準區域」係為圖1〇A中之位於 腫瘤輪廓外側（即位於腫瘤外部區域）之矩形框所包圍的區 域。一般而言，此「基準區域」係位於代表正常組織的影 像區域内。 在本實施例中’迴音性特徵係藉由將位於此腫瘤内部 區域内之此等|素點戶斤具有《灰階梯度值的平均值“為) 減去位於此基準區域内之此等像素點所具有之灰階梯度值 201126260 的平均值（wtw,c?,.)所得到之差值除以位於此基準區域内之此 等像素點所具有之灰階梯度值的平均值的方式被量 化，即如下列式子所示：2Q 201126260 In addition, in the imaging method of the tumor edge feature of the present embodiment, the rainbow color gradation system matched by the step (E) is a continuous gradation gradation of a red, yellow, green and blue drum purple. When the tumor edge features are imaged, each pixel point located on the measurement line segment is imaged according to the following imaging rules: 1 · If the pixel located on the measurement line segment has If the gray-scale moving variation value is greater than or equal to the above-mentioned "edge imaging upper bound", the pixel is covered by a red block; 2. If the pixel located on the measuring line segment has a gray-scale moving variation value smaller than Or equal to the aforementioned "edge imaging lower bound", the pixel is covered by a purple block; and 3. If the pixel on the measuring line segment has a gray-scale moving variation value between the aforementioned "edge imaging" The upper bound" and the aforementioned "edge imaging lower bound" are based on the correspondence between the gray-scale moving variation values of the pixel points and the aforementioned "edge imaging upper bound" and the aforementioned "edge imaging lower bound" respectively. The pixel is covered by a block having a color corresponding to the rainbow level. After the step (E) is completed, the imaging method of the tumor edge feature of the embodiment may further include a step (F) after the step (E) is completed by imaging the pixels on the measurement line segment according to the above imaging rule. Change the position of this section line to sweep all the edges of the tumor. In this way, all pixels in the annular region of the tumor contour on all edges of the tumor can be imaged to obtain a tumor edge feature image as shown in FIG. The distribution of the edge features of the tumor and the degree of blurring of the tumor edge can be easily judged by the imaging of the tumor edge feature as shown in Fig. 5 of Fig. 5 201126260. ... It should be noted that when the tumor edge features are imaged, the color blocks obtained by matching the shades of the shirts are continuously covered with the gray P&~ images to display these colors intermittently. On this grayscale image of the block, Dr. Eli observed the other features of the tumor at the same time. Fig. 6A is a schematic diagram of an ultrasonic grayscale image, which is composed of a plurality of pixel points, each having a gray gradient value. As shown in Fig. 6A, it shows a thyroid tumor and its surrounding thyroid tissue, and this thyroid tumor contains a cyst area. 6B is a flow chart of a method for quantifying tumor cyst characteristics according to a third embodiment of the present invention. The method includes the following steps: (A) performing a tumor contour and a tumor contour annular region from the gray scale image, and the tumor contour system is Located in the annular region of the tumor contour; (B) overlaying the contour of the tumor on the grayscale image to define an internal region of the tumor and an external region of the tumor on the grayscale image; (C) by being located here The gray gradient values of the pixels in the inner region of the tumor respectively, and the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor are calculated; And (D) quantifying the cyst feature 1 located in the inner region of the tumor based on the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor. 201126260 Among them, the "tumor contour" and "tumor contour contour region" taken out in step (A) and the "tumor inner region" and "external tumor region" defined by "tumor contour" in step (B) It has been described in detail in the month and will not be repeated here. In addition, since the method of calculating the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the "inside of the tumor" in step (C) has been widely known, This will not be repeated. In addition, as shown in FIG. 6C, in the method for quantifying tumor cyst features of the present embodiment, step (D) includes a step (D1) by which the pixels located in the inner region of the tumor have The minimum value of the gray gradient value and the standard deviation of the gray gradient value define a threshold value of the gray gradient value of the cyst feature to calculate the cyst characteristic in the inner region of the tumor. proportion. In this embodiment, the threshold is the minimum value (heart ^/) of the gray gradient value (GV/) of the pixels located in the inner region of the tumor plus the zero point of the gray The standard deviation of the step value is '丨χ'. If a gray level value (G〃/) of such a pixel in the inner region of the tumor is below this threshold, the pixel is defined as having a cyst characteristic. After comparing all the pixels located in the inner region of the tumor, the number of pixels defined as having cyst characteristics divided by the number of all pixels located in the inner region of the tumor can be used to obtain cyst characteristics. The proportion of the area inside the tumor. 7A is a flow chart of a method for imaging a tumor cyst feature according to a fourth embodiment of the present invention, comprising the following steps: 32 201126260 (A) Obtaining a tumor contour and a tumor wheel corridor annular region from the grayscale image, and The tumor contour is located in the annular region of the tumor wheel; (B) The tumor contour is superimposed on the gray scale image to define an internal tumor region and an external tumor region on the gray scale image: (C Calculating the minimum value of the gray gradient value and the gray gradient of the pixels located in the inner region of the tumor by the gray gradient values respectively obtained by the pixels located in the inner region of the tumor The standard deviation of the values; and (D) define a cyst imaging upper bound and a cyst based on the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor. The lower bound is imaged to visualize the characteristics of the tumor cyst located within the interior region of the tumor. Among them, the "tumor contour" and the "tumor contour loop" area extracted from the step (A) and the "tumor area" and "tumor outside" defined by the "tumor contour" in the step (B) The area has been described in detail in the 'before' and will not be repeated here. In addition, the method of calculating the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the "inside of the tumor" due to the step (C) has been widely known. This will not be repeated. In addition, in the imaging method of the tumor cyst feature of the present embodiment, the upper boundary of the capsule imaging of the step (D) is the gray gradient value of the pixels located in the inner region of the tumor. The minimum value (w, .„Gw) plus zero-times the standard deviation of the gray gradient value (4), ie 4/+0.丨X.4. 201126260 On the other hand, the lower bound of cyst imaging is located The minimum value of the gray gradient value (^) of the pixels in the inner region of the tumor. It should be noted that, under other application conditions, the aforementioned "cyst imaging upper bound" and "cyst imaging lower bound" It may also have different values, such as "the upper boundary of the cyst imaging", which may be the minimum value of the gray gradient value of the pixels located in the inner region of the tumor plus the value of the gray gradient value of three times three times. The standard deviation, (ie, m"7(^+0.3Xi,〆^), "the lower boundary of the cyst imaging" can add zero to the minimum value of the gray gradient value of the pixels located in the inner region of the tumor. Point zero or five times the standard deviation of this gray gradient value (ie, as long as "the cyst is imaged The value of the boundary is greater than the value of the lower boundary of the cyst imaging. In addition, when the tumor cyst features are imaged, if the pixel located in the inner region of the tumor has a gray gradient value is in the foregoing Between the upper boundary of the cyst imaging and the lower boundary of the cyst imaging, the pixel is covered by the pink patch. Thus, after each pixel located in the inner region of the tumor is visualized according to the rule, The image of the tumor cyst features as shown in Fig. 7B can be obtained, and by visualizing the image of the tumor cyst as shown in Fig. 7B, the physician can easily judge the distribution of the cyst characteristics in the tumor. And the ratio of the inner region of the tumor. In addition, the imaging method of the tumor cyst feature of the fourth embodiment of the present invention further comprises the following steps after the step (D): (E) the inner region of the tumor is located The equal pixel points are respectively defined as a plurality of reference masks, and each of the reference masks includes a reference pixel point and a plurality of pixel points adjacent to the reference pixel point 'and the reference image 34 201126 The gray gradient value of the 260 prime point is between the lower bounds of the image; and the upper boundary of the cyst image and the cyst (less) when at least the silk is removed; , etc. = : When the image is between the lower boundary of the tumor, the reference pixel point and the pixel are swollen (4). The material of the impurity is covered by the pink block. The shadow of the tumor cyst is shown in the fourth embodiment. The method further comprises the step (g) after the step (F) described above, when only the gray pixel value of the reference pixel point is between the upper boundary of the cyst imaging and the lower boundary of the metastatic cyst image is removed Covering the reference pixel and the pink block of the pixels. μ Thus, by completing the foregoing steps (Ε) to (G), the shape and area of the pink block and the actual tumor characteristics The shape and actual area are better matched. Further, in the present embodiment, the reference mask defined in the step (E) contains 9 pixel points. Fig. 8A is a schematic diagram of an ultrasonic grayscale image, which is composed of a plurality of pixel points, and each pixel has a gray gradient value. As shown in Fig. 8A, it shows a squamous gland tumor and its surrounding sacral gland tissue, and this thyroid tumor contains an I-segmented region. 8B is a flow chart of a method for quantifying a tumor maturation feature according to a fifth embodiment of the present invention. The method includes the following steps: (A) extracting a tumor contour and a tumor contour annular region from the grayscale image and the tumor contour system Located in the annular region of the tumor contour; 201126260 (B) overlaying the contour of the tumor on the grayscale image to define an internal region of the tumor and an external region of the tumor on the grayscale image; (C) by The gray level values of the pixels in the inner region of the tumor respectively, and the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor are calculated. (D) extracting one of the grayscale images from the grayscale image based on the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor The cyst area; (E) calculated by the gray gradient value of the pixels located in the inner region of the tumor but outside the cyst region, respectively The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels in the domain but outside the cyst area; and (F) the internal region of the tumor The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels located outside the cystic region, which will be located in the tumor internal calcification Feature quantization. Among them, the "tumor contour" and the "tumor contour annular region" taken out in the step (A) and the "tumor inner region" and "external tumor region" defined in the "tumor contour" in the step (B) The "cyst area" taken in step (D) has been described in detail above, and will not be repeated here. In addition, the method of calculating the minimum value of the gray gradient value and the standard deviation of the gray gradient value 201126260 of the pixels located in the "inside of the tumor" due to the calculation of the step (C), and the calculation of the step (E) The method of presenting the maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels located in the "inside of the tumor" but outside the "cyst area" has been widely used. It is familiar to all walks of life' and will not be repeated here. In addition, as shown in FIG. 8C, 'in the method for quantifying the calcification characteristics of the tumor of the present embodiment', the step (F) includes a step (F1) by being located in the inner region of the tumor but outside the cyst region. The maximum value of the gray gradient value (G)/) of these pixels, the standard deviation of the gray gradient value and the average value of the gray gradient value (we(w, defining the gray gradient of a calcification feature) Thresho丨d value of the value to calculate the proportion of this calcification in the internal region of the tumor. In this example, the threshold is located in the internal region of the tumor but is located in the cyst. The average value of the gray gradient value (g,) of such pixels outside the region (plus a factor of eight times the standard deviation of the gray gradient value («<_ and /), ie coffee - Horse +2·8Χ_(7〆If the gray gradient value of such a pixel in the inner region of the tumor but outside the cyst region is higher than the threshold, the pixel is defined as having Calcification characteristics. After aligning all the areas located inside the tumor but outside the cyst area After the pixel, after dividing the number of pixels with calcification characteristics by the number of all pixels located in the inner region of the tumor, the proportion of calcification in the inner region of the tumor can be obtained. 9A is a flow chart of a method for imaging a tumor calcification characteristic according to a sixth embodiment of the present invention, comprising the following steps: 201126260 (A) extracting a tumor contour and a tumor wheel _ jade-shaped region from the gray-scale image, and The contour of the tumor is located in the annular region of the tumor contour; (B) the contour of the tumor is superimposed and displayed on the grayscale image to define an internal region of the tumor and an external region of the tumor on the grayscale image; (C) Calculating the minimum value of the gray gradient value and the gray gradient value of the pixels located in the inner region of the tumor, respectively, from the gray gradient values of the pixels located in the inner region of the tumor. Standard deviation; (D) From the grayscale image, the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor a cyst area within the inner region of the tumor; (E) calculated by the gray gradient value of each of the pixels located in the inner region of the tumor but outside the cyst region The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels in the region but outside the cyst region; and (F) based on the inside of the tumor The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels in the region but outside the cyst region define an upper bound of the calcification image and The lower limit of calcification imaging is to visualize the calcification features of the tumor located in the inner region of the tumor. Among them, the "tumor contour" and the "tumor contour annular region" taken out in step (A) and the step (B) are used. The "2011 201126260 Tumor Area" and "External Tumor Area" and the "Cell Area" extracted in Step (D) have been described in detail before, and will not be described here. Repeat the remarks. In addition, the method (step) (c) calculates the minimum value of the gray gradient value and the private standard deviation of the gray gradient value of the pixels located in the "inside of the tumor". The method of calculating the maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value of the pixels located in the "intratumoral region" but outside the "cyst region" has been calculated. It is widely known to all walks of life' and will not be repeated here. In addition, in the imaging method of the tumor calcification feature of the present embodiment, the calcified imaging upper boundary of step (F) is such that the pixels located in the inner region of the tumor but outside the cyst region have The maximum value of the gray gradient value, and the lower bound of the imaginary imaging is the average value of the gray gradient value of the pixels located in the inner region of the tumor but outside the cyst region plus two to eight times The standard deviation of this gray gradient value is 顺^[(7,)/+2.8. It should be noted that under other application conditions, the above-mentioned "calcification imaging upper bound" and "calcification imaging lower bound" may also have different values, such as "calcification imaging upper bound" may be located in the inner region of the tumor but located The maximum value of the gray gradient value of the pixels outside the cyst area minus the zero value of the standard deviation of the gray gradient value, that is, the lower bound of the image formation of _〇" ' The average value of the gray gradient values of the pixels located in the inner region of the tumor but outside the cyst region (mra plus two-fifths the standard deviation of the gray gradient value 201126260 (i./ i/a cO(y/), ie mean c +2.5 Xs, heart.C^y/ 'As long as the value of the "upper bound upper bound" is greater than the value of the "calculus lower bound". In addition, when the tumor is calcified During imaging, if the pixel located in the inner region of the tumor but outside the cyst region has a gray gradient value between the above-mentioned "calcification imaging upper boundary" and "calcification imaging lower boundary", This pixel is covered by a yellow block. Therefore, after each pixel located in the inner region of the tumor but outside the cyst region is visualized according to the rule, the tumorized image of the tumor as shown in FIG. 9B can be obtained. By visualizing the image of the tumor reduction feature as shown in Fig. 9B, the physician can easily judge the distribution of the tumor in the tumor and the ratio of the inner region of the tumor. Fig. 10A is an ultrasound A schematic diagram of a grayscale image, which is composed of a plurality of pixel points, and each pixel has a gray gradient value. As shown in Fig. 10A, it shows a thyroid tumor and its surrounding thyroid tissue. A flow chart of a method for quantifying tumor echogenicity according to a seventh embodiment of the present invention, comprising the steps of: (A) operating a gray contour image from a gray contour image and a tumor contour annular region, and the tumor contour is located here (B) overlaying the contour of the tumor on the grayscale image to define an internal region of the tumor and an external region of the tumor on the grayscale image; (C) The gray gradient values of the pixels in the inner region of the tumor respectively calculate the average value of the gray gradient values of the pixels located in the inner region of the tumor; 40 201126260 (9) The outer region selects the #_reference region, and the gray gradient values of the pixels located in the reference region respectively have the gray gradient values of the pixels located in the reference region: The average value; and (6) the average value of the gray gradient values of the pixels located in the inner region of the tumor and the average value of the gray gradient values of the pixels located in the reference region The echogenic feature of this meaning is quantified. Among them, the "tumor contour" and the "tumor contour annular region" taken out in step (4) and the "tumor inner region" and "the tumor" defined in step (B). The outer region of the tumor has been described in detail before, and will not be repeated here. In addition, the method of calculating the average value of the gray gradient values of the pixels located in the "inside of the tumor" in step (C) and the calculation of the step (D) in the reference region The method of averaging the gray gradient values of the pixels has been widely known and will not be described here. In addition, the aforementioned "reference area" is an area surrounded by a rectangular frame located outside the tumor outline (i.e., located outside the tumor area) in Fig. 1A. In general, this “reference area” is located in the image area representing normal tissue. In the present embodiment, the 'echo feature' is obtained by subtracting the pixels located in the reference region from the "average value of the gray gradient value" of the pixels located in the inner region of the tumor. The difference between the average value (wtw, c?, .) of the gray gradient value 201126260 of the point is divided by the average value of the gray gradient values of the pixels located in the reference area. Quantification, as shown in the following formula:

ER ^ean jjj mean mean xlOO% 由上式計算所得的結果，一般以ER表示。當ER大於或 等於零時（即^ 〇)，此腫瘤便具有高迴音性（Hyperechoic) 之特徵。另一方面，當ER小於零時（即£/?<〇)，此腫瘤便具 有低迴音性（Hyp0echoic)之特徵。 圖11A係一超音波灰階影像的示意圖，其係由複數個 像素點組合而成，且每一像素點分別具有一灰階梯度值。 如圖11A所示’其顯示一曱狀腺腫瘤與其周圍的甲狀腺組 織》 圖1 1 B係本發明第八實施例之腫瘤異質化特徵之量化 方法的流程圖，其包括下列步驟： (A) 從此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廊環 形區域’且此腫瘤輪廓係位於此腫瘤輪廓環形區域内； (B) 將此腫瘤輪廓重疊顯示於此灰階影像上，以在此灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 將位於此腫瘤内部區域内之此等像素點分別定義 為複數個參考遮罩’且每一此等參考遮罩係包含一基準像 素點與複數個相鄰於此基準像素點之像素點； 42 201126260 ⑼計算出每-此等參考遮罩所分別具有之參考遮罩 灰階梯度值局部平均（丨⑽丨_翁參考料灰階悌度值局部 變異（local variance); (E) 汁异出每__此等參考遮罩所分別具有之參考遮罩 灰階梯度值局部平均之變異（variance 〇f _ mean)、來考遮罩 灰階梯度值局部變異之平均（_ Gf丨⑽丨__)以及來考遮 罩灰階梯度值局部變異之變異（variance 〇f —咖叫；以及 (F) 藉由每一此等參考遮罩所分別具有之至少—選自 於-由參考遮罩灰階梯度值局部平均之變異參考遮罩灰 階梯度值局部變異之平均以及參考遮罩灰階梯度值局部變 異之變異所構成之群組，計算出每—此等參考遮罩所分別 具有之異質化指標值，將此腫瘤之異質化特徵量化。 其中，由於步驟（A)所擷取出之「腫瘤輪廓」與「腫瘤 輪廓環形區域」以及步驟（B)之藉由「腫瘤㈣」所定義出 之「腫瘤内部區域」與「腫瘤外部區域」均已詳細敛述於 前’在此便不再重複敘述。 此外，在本實施例中，步驟（c)之參考遮罩係包含乃個 像素點，即基準像素點以及與複數個相鄰於此基準像素點 的像素點，而這些像素點之座標值（x，，y，）與此基準像素點之 座標值（x，y)之間的差距均在2以内。需注意的是，在其他的 應用狀況下，步驟（C)之參考遮罩亦可包含不同數目的像素 點，如49個像素點。 以下，依據每一此等參考遮罩所分別具有之參考遮罩 灰階梯度值局部平均（local mean)及參考遮罩灰階梯度值局部 43 201126260 變異（丨oca丨variance)而計算出每一此等參考遮罩所分別具有之 v考k罩灰匕梯度值局部平均之變異（var丨丨〇cai mean, VOM)、參考遮罩灰階梯度值局部變異之平均（mean 〇n〇cal variance, MOV)以及參考遮罩灰階梯度值局部變異之變異 (variance of local variance，v〇v)的過程，將詳細敘述。 首先，參考遮罩灰階梯度值局部平均之變異（v〇M)可被 解讀為局# +均的變化。意' #，參考遮罩灰階梯度值局部 平均之變異（VOM)可被計算為存在於各局部平均之間的變 異其-人，參考遮罩灰階梯度值局部變異之平均可被 解讀為局部區間（l〇cal area)的差異幅度。意即，參考遮罩灰 階梯度值局部變異之平均（M〇v)可被計算為存在於各局部 區間（local area)變異之間的平均。最後，參考遮罩灰階梯度 值局部變異之變異（VOV)可被解讀為存在於各局部區間之 間之差異的變化意即，參考遮罩灰階梯度值局部變異之變 異（VOV)可被計算為存在於各局部區間（1〇cal area)變異之間 的變異。 一用於解釋這三種變異之計算的例子，則提供於下： 首先，針對一位於一超音波灰階影像中的一量測點 Pi，j，一局部矩形區域（如參考遮罩）被定義出來（表示為 心），其包含（2q+l)x(2q+l)個量測點，其中q係用於 定義矩形區域寬度的參數。此外，的組成元素可表示 44 201126260 P.. p.. • · • · … Pi+qj~l! » • ♦ P . « • • · …piJ • · • · • • « • P 卜 q'J+q * · p.. •.. Pi+q-.i+9ER ^ean jjj mean mean xlOO% The result calculated by the above formula, generally expressed as ER. When ER is greater than or equal to zero (i.e., ^ 〇), the tumor is characterized by a hyperechoic. On the other hand, when ER is less than zero (i.e., £/?<〇), the tumor is characterized by a low echogenicity (HypOechoic). Figure 11A is a schematic diagram of an ultrasonic grayscale image, which is composed of a plurality of pixel points, and each pixel has a gray gradient value. As shown in FIG. 11A, 'it shows a squamous gland tumor and its surrounding thyroid tissue. FIG. 11B is a flow chart of a method for quantifying tumor heterogeneity characteristics according to an eighth embodiment of the present invention, which comprises the following steps: (A) From this grayscale image, a tumor contour and a tumor circle annular region are taken out and the tumor contour is located in the annular contour region of the tumor contour; (B) the tumor contour is superimposed and displayed on the grayscale image to be grayed out The inner image of the tumor and the outer region of the tumor are defined on the order image; (C) the pixels located in the inner region of the tumor are respectively defined as a plurality of reference masks and each of the reference masks includes one a reference pixel point and a plurality of pixel points adjacent to the reference pixel point; 42 201126260 (9) Calculating a local average of the reference mask gray gradient value for each of the reference masks (丨(10)丨_翁 reference material Gray-scale 值 值 local variation ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Test cover The average of the local variation of the gray gradient value (_ Gf丨(10)丨__) and the variation of the local variation of the gray scale value of the mask (variance 〇f - coffee call; and (F) by each such reference The hood has at least - selected from the group consisting of - the average of the local average variation of the reference mask gray gradient value, the average of the local variation of the reference mask gray gradient value, and the variation of the local variation of the reference mask gray gradient value The group calculates the heterogeneity index value of each of the reference masks, and quantifies the heterogeneous characteristics of the tumor. Among them, the "tumor contour" and the "tumor contour annular region" taken out in step (A) And the "inside of the tumor" and the "outside of the tumor" defined by the "tumor (4)" in step (B) have been described in detail in the previous section, and the description will not be repeated here. In addition, in this embodiment The reference mask of the step (c) includes a pixel point, that is, a reference pixel point and a plurality of pixel points adjacent to the reference pixel point, and the coordinate values of the pixel points (x, y,) With this reference pixel The difference between the coordinate values (x, y) is within 2. It should be noted that in other applications, the reference mask of step (C) may also contain different numbers of pixels, such as 49 pixels. The following is calculated based on the local mean of the reference mask gray value and the reference mask gray gradient value local 43 201126260 variation (丨oca丨variance) for each of the reference masks. The average of the var丨丨〇cai mean (VOM) and the local variation of the reference mask gray gradient value for each of the reference masks (mean 〇n〇) Cal variance, MOV) and the process of variation of local variance (v〇v) of reference mask gray gradient values will be described in detail. First, the variation of the local average of the reference grayscale value (v〇M) can be interpreted as the change of the locality.义 '#, reference mask gray gradient value local average variation (VOM) can be calculated as the variation between the local averages - the average of the local variability of the reference mask gray gradient value can be interpreted as The difference in the local interval (l〇cal area). That is, the average (M〇v) of the local variability of the reference mask gray gradient value can be calculated as the average existing between the local area variations. Finally, the variation of the local variation of the reference gray scale value (VOV) can be interpreted as the change in the difference between the local intervals, that is, the variation of the local variation of the reference gray scale value (VOV) can be It is calculated as the variation existing between the variations of the local area (1〇cal area). An example for explaining the calculation of these three variations is provided below: First, for a measurement point Pi,j located in an ultrasonic grayscale image, a partial rectangular region (such as a reference mask) is defined. Coming out (denoted as a heart), which contains (2q + 1) x (2q + 1) measurement points, where q is used to define the parameters of the width of the rectangular area. In addition, the constituent elements can represent 44 201126260 P.. p.. • · • · ... Pi+qj~l! » • ♦ P . « • • · ...piJ • · • • • • « • P Bu q'J +q * · p.. •.. Pi+q-.i+9

Alj 乂<7 位於心内之各點的樣本平均則被定義為局部平均 μ2 (local mean)，其可表示為 ，且可以下式計算：The average of the samples of Alj 乂<7 at each point in the heart is defined as the local mean μ2 (local mean), which can be expressed as , and can be calculated as:

其中’ j係Pi,j於― (frame)的亮度。 位於以内之各點的 (local variance) ’ 其可表示為 固定時間點（timepoint)之單—架構 樣本變化則被定義為局部變異 〆 、 巧^· ’且可以下式計算： 201126260 i+cl j+q 以 瘤 示Where 'j is Pi, j is the brightness of ― (frame). The local variance of the points (which can be expressed as a fixed time point) - the architectural sample change is defined as the local variation 〆, Qiao ^· ' and can be calculated as: 201126260 i+cl j +q to show

S A(!; 1 一q j =j-q (2g + l)2 -1S A(!; 1 a q j =j-q (2g + l)2 -1

Aq. 然而’當 包含位於有興趣區域（region of interest, ROI)μ2 q s2q 外的點時，以的 及<·將不會被計算。 基於先前定義出之有興趣區域（R0I)，在此例子中，腫 内部區域，且q具有一定值，即q=2。而、局部平均可表 a2 2 為〜局部變異則可表示為其中1<v< 因 此，參考遮罩灰階梯度值局部平均之變異（v〇M)可zk-s " V〇Ma =— Η 其中 η. Σ> 2 ν- η q 46 201126260 參考遮罩灰階梯度值局部變異之平均（MOV)則可計算 為：Aq. However, when a point other than the region of interest (ROI) μ2 q s2q is included, <· will not be calculated. Based on the previously defined region of interest (R0I), in this example, the inner region is swollen, and q has a certain value, i.e., q=2. However, the local average can be a2 2 to ~ local variation can be expressed as 1 < v < Therefore, the reference mask gray gradient value local average variation (v〇M) can be zk-s " V〇Ma = Η where η. Σ> 2 ν- η q 46 201126260 The mean gradation (MOV) of the reference mask gray gradient value can be calculated as:

最後，參考遮罩灰階梯度值局部變異之變異（VOV)則可 計算為Finally, the variation of the local variation of the reference gray gradient value (VOV) can be calculated as

其中， 2Of which 2

η 5* 201126260 =，參考Μ灰階梯度值局料均之變異（v 考遮罩灰階梯度值局部變異之平均⑽加及參考遮罩灰 階梯度值局部變異之變異（vqV)的計算過程以及 灰階梯度值局部變異之變異(vov)除以參考遮罩灰階梯戶 ==均之變異(V0M)所得之比值、參考遮罩灰階梯度： 局錢異之平均（M0V)除以參考遮罩灰階梯度值局部變異 ^異（卿)所得之比值以及參考遮罩灰階梯度值局部變 異之平均_V)除以參考遮罩灰階梯度值局部㈣ (VOM)所得之比值的計算過程，也在此被揭露。 、 在前述之6種形式的異質化指標值中，其中當異質化指 標值係係一變數為參考遮罩灰階梯度值局部變異之變^ (VOV)除以參考遮罩錢梯度值局部平均之 之比值的函數時，此異質化指標值具有最佳的表(現”付 圖12A係本發明第九實施例之腫瘤異質化特徵之影像 化方法的流程圖，其包括下列步驟： ,(A)從此灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環 形區域，且此腫瘤輪廓係位於此腫瘤輪廓環形區域内； (B) 將此腫瘤輪廓重疊顯示於此灰階影像上以在此灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 將位於此腫瘤内部區域内之此等像素點分別定義 為设數個參考遮罩，且每一此等參考遮罩係包含一基準像 素點與複數個相鄰於此基準像素點之像素點； 48 201126260 (D) 藉由包含於每_卜 別具有之灰階梯度值，計算寺出每考4罩…像素點所分 有之參考遮罩讀梯度值變異；*" ^考遮罩所分別具 (E) 藉由每一此等參老沒 考〜罩所刀別具有之參考遮罩灰 1¾梯度值變異，計算出此箄來 階梯度值變異； ^考遮罩所具有之平均遮罩灰 (F) 措由母一此寺參考遮罩所分別具有之 階梯度值變異及此等參考避目+ τ 早Λ 靜里η- 所具有之平均遮罩灰階梯度 值變異’计异出母一此等夂去、毋 寺，考遮罩所分別具有之異質化指 標值； (G) 錯由母一此等參考砂蓄糾八。，Η > 可遮罩所分別具有之異質化指標 值，計算出此等參考遮罩所 可圮卓所具有之異質化指標值的最大 值、異質化指標值的最小值、異質化指標值的平均值及里 質化指標值的標準差；以及 (H) 依據此等參考遮罩所具有之異質化指標值的最大 值、異質化指標值的最小值、異質化指標值的平均值及異 質化指標值的標準差，定義出-異質化成像上界及—異質 化成像下界’且配合-彩虹色階將位於此腫瘤内部區域内 之腫瘤異質化特徵影像化。 其中，由於步驟（A)所擷取出之「腫瘤輪廓」與「腫瘤 輪廓環形區域」以及步驟（B)之藉由「腫瘤輪廓」所定義出 之「腫瘤内部區域」與「腫瘤外部區域」均已詳細敘述於 前，在此便不再重複敘述。 ' 49 201126260 此外，由於步驟（G)之計算出此等參考遮罩所具有之異 質化指標值之最大值、異質化指標值之最小值、異質化指 標值之平均值及異質化指標值之標準差的方法已廣為各界 所熟悉’在此便不再贅述。 除此之外，在本實施例中，步驟（c)之參考遮罩係包含 25個像素點，即基準像素點以及與複數個相鄰於此基準像 素點的像素點，而這些像素點之座標值（χ，，ν，）與此基準像素 點之座標值（x，y)之間的差距均在2以内。需注意的是，在其 他的應用狀況下，步驟（C)之參考遮罩亦可包含不同數目的 像素點’如49個像素點。 除此之外，前述之步驟（D)所計算出之每一此等參考遮 罩（M//4)所分別具有的「參考遮罩灰階梯度值變異」係以 、％#///,.表示，步驟（E)所計算出之此等參考遮罩所具有之「平 均遮罩灰階梯度值變異」則以λ,〇κΜ//，表示，且 Σ^ρΜΗΙι:η 5* 201126260 =, refer to the variation of the grading value of the ash grading value (v measur the average of the local variability of the gray scale value of the mask (10) plus the variation of the local variability of the reference mask gray gradient value (vqV) And the variation of the local variation of the gray gradient value (vov) divided by the ratio of the reference mask gray step household == variability (V0M), the reference mask gray gradient: the average of the money difference (M0V) divided by the reference The ratio of the ratio of the local variability of the mask gray gradient value to the average value of the local variability of the reference mask gray gradient value _V) divided by the reference mask gray gradient value local (four) (VOM) The process is also revealed here. Among the six types of heterogeneity index values mentioned above, where the heterogeneity index value is a variable is the reference mask gray gradient value local variation variation (VOV) divided by the reference mask money gradient value local average In the case of a ratio of the values, the heterogeneity index value has the best table (now "Figure 12A" is a flowchart of the imaging method of the tumor heterogeneity feature of the ninth embodiment of the present invention, which includes the following steps: A) extracting a tumor contour and a tumor contour annular region from the grayscale image, and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the grayscale image to be grayed out The inner region of the tumor and the outer region of the tumor are defined on the order image; (C) the pixels located in the inner region of the tumor are respectively defined as a plurality of reference masks, and each of the reference masks includes a reference pixel point and a plurality of pixel points adjacent to the reference pixel point; 48 201126260 (D) Calculated by the gray gradient value included in each _b, the temple is calculated 4 points per pixel... Reference mask Read gradient value variation; *" ^ test masks have (E) by each of these ginsengs have not tested ~ cover knives have reference mask gray 13⁄4 gradient value variation, calculate this step The variation of the degree value; the average mask gray (F) of the test mask has the gradient value variation of the reference mask of the mother, and the reference avoidance + τ early 静 静 η- The average mask gray gradient value variation has the same heterogeneous index value as the mask, and the mask has the heterogeneous index value; (G) the wrong reference to the sand ., Η > can mask the heterogeneity index values, calculate the maximum value of the heterogeneous index value, the minimum value of the heterogeneity index value, and the heterogeneity index of the reference mask. The average value of the value and the standard deviation of the grading index value; and (H) the maximum value of the heterogeneous index value, the minimum value of the heterogeneity index value, and the average value of the heterogeneity index value according to the reference masks And the standard deviation of the heterogeneity index values, defining - the heterogeneous imaging upper bound and - heterogeneous The lower bound 'and the fit - rainbow scale will visualize the tumor heterogeneity features located in the inner region of the tumor. Among them, the "tumor contour" and "tumor contour annular region" taken out in step (A) and step (B) The "intratumoral area" and "external tumor area" defined by "tumor contour" have been described in detail before, and will not be repeated here. ' 49 201126260 In addition, due to the calculation of step (G) The method of determining the maximum value of the heterogeneous index value, the minimum value of the heterogeneity index value, the average value of the heterogeneity index value, and the standard deviation of the heterogeneity index value of these reference masks is widely known. In addition, in this embodiment, the reference mask of step (c) includes 25 pixel points, that is, a reference pixel point and a plurality of pixel points adjacent to the reference pixel point, The difference between the coordinate values (χ, ν,) of these pixels and the coordinate value (x, y) of the reference pixel is within 2. It should be noted that in other applications, the reference mask of step (C) may also contain a different number of pixels, such as 49 pixels. In addition, the "reference mask gray gradient value variation" of each of the reference masks (M//4) calculated in the foregoing step (D) is, %#/// , indicating that the "average mask gray gradient value variation" of the reference masks calculated in step (E) is represented by λ, 〇κΜ//, and Σ^ρΜΗΙι:

MoiMHIj = - z' 另方面如圖丨1 c所示，在本實施例之腫瘤異質化特 徵的量化方法中’步驟（F)包含一步驟（F1)，藉由將此參考 遮罩所具有之參考遮罩灰階梯度值變異。声所,)減去此等 多考遮罩所具有之平均遮罩灰階梯度值變異（⑽，娜/,)所得 之差值的絕對值開根號的方式，計算出每一此等參考遮罩 所分別具有之異質化指標值斤户，，如下列式子所示： 201126260 再如圖1丨c所示’在本實施例之腫瘤異質化特徵的量化 方法中’步驟（F)於步驟（F1)之後更包括一步驟（F2)，藉由每 一此等參考遮罩所分別具有之異質化指標值（//P,·:)，定義出 異貝化特徵之異質化指標值的閥值(threshold value)，以計算 出此異質化特徵於此腫瘤内部區域内所佔的比例。 在本實施例之腫瘤異質化特徵的影像化方法中，步驟 (H)之異質化成像上界係為此等參考遮罩所具有之異質化 指標值（///\-)的最大值’異質化成像下界則為此等參考遮罩 所具有之異質化指標值的平均值減去零點一倍之異 質化指標值的標準差（^户,），即。 需注意的是’在其他的應用狀況下，前述之「異質化 成像上界」及「異質化成像下界」亦可具有不同的數值， 如「異質化成像上界」可為異質化指標值（///\)的最大值減 去零點一倍之異質化指標值的標準差,「異質化成像 下界」則可為此等參考遮罩所具有之異質化指標值的平均 值（mea/W3,.)減去零點二倍之異質化指標值的標準差 ，只要「異質化成像上界」的數值大 於「異質化成像下界」的數值即可。 此外’在本實施例之腫瘤異質化特徵的影像化方法 中，步驟（H)所配合的此彩虹色階係為一紅橙黃綠藍靛紫之 連續漸變色階。而在腫瘤異質化特徵影像化時，每一此等 參考遮罩所具有炙基準像素點係分別依據下述之影像化規 則而被影像化： 201126260 •右此參考遮罩所具有之異質化指標值等於前述之 「 異質化成像上界」，則以一红色區塊覆蓋此參考 遮罩之基準像素點； 2. 若此參考遮罩所具有之異質化指標值小於或等於 前述之「異質化成像下界」，則以一紫色區塊覆蓋 此參考遮罩之基準像素點；以及 3. 右此參考遮罩所具有之異質化指標值介於前述之 「異質化成像上界」與前述之「異質化成像下界」 之間’便依據此參考遮罩所具有之異質化指標值分 別與前述之「異質化成像上界」與前述之「異質化 成像下界」之間的對應關係，以一具有從此彩虹色 階中對應出之顏色的區塊覆蓋此參考遮罩之基準 像素點。 之基準像素點影像化之後 化特徵影像化圖像。而藉1 影像化圖像，醫師可鉬县 而在依據上述影像化規則將每一此等參考遮罩所具有 ’即付出如圖12B所示之腫瘤異質MoiMHIj = - z' In another aspect, as shown in Fig. 1c, in the method for quantifying tumor heterogeneity characteristics of the present embodiment, 'step (F) comprises a step (F1) by which the reference mask has Refer to the mask gray gradient value variation. Acoustic,) subtract each absolute reference from the absolute value of the difference in the mean mask gray gradient value ((10), Na/,) of these multiple masks. The mask has a heterogeneous index value, respectively, as shown in the following formula: 201126260 As shown in Figure 1丨c, 'in the quantification method of tumor heterogeneity characteristics in this embodiment' step (F) Step (F1) further includes a step (F2) of defining a heterogeneous index value of the heterobeat feature by each of the reference masks having a heterogeneous index value (//P, ·:) Threshold value to calculate the proportion of this heterogeneous feature in the internal region of the tumor. In the imaging method of the tumor heterogeneity feature of the present embodiment, the heterogeneous imaging upper bound of step (H) is the maximum value of the heterogeneous index value (///\-) of the reference mask for this reference. The lower bound of the heterogeneous imaging is the standard deviation of the heterogeneous index value (^ household,), which is the average of the heterogeneous index values of the reference mask. It should be noted that in other applications, the above-mentioned "heterogeneous imaging upper bound" and "heterogeneous imaging lower bound" may also have different values, such as "heterogeneous imaging upper bound" may be a heterogeneous index value ( The maximum value of ///\) is subtracted from the standard deviation of the zero-point heterogeneity index value. The "heterogeneous imaging lower bound" can be the average value of the heterogeneous index values of the reference mask (mea/ W3,.) minus the standard deviation of the zero-fold heterogeneity index value, as long as the value of the "heterogeneous imaging upper bound" is larger than the value of the "heterogeneous imaging lower bound". Further, in the imaging method of the tumor heterogeneity feature of the present embodiment, the rainbow gradation system matched by the step (H) is a continuous gradation of red orange, yellow, green, blue and purple. When the tumor heterogeneity features are imaged, the reference pixel points of each of the reference masks are respectively imaged according to the following imaging rules: 201126260 • The heterogeneity index of the reference mask is right The value is equal to the above-mentioned "heterogeneous imaging upper bound", and the reference pixel of the reference mask is covered by a red block; 2. If the reference mask has a heterogeneous index value less than or equal to the aforementioned "heterogeneization" The lower bound of the image is covered by a purple block covering the reference pixel of the reference mask; and 3. The right heterogeneous index value of the reference mask is between the aforementioned "heterogeneous imaging upper bound" and the aforementioned " The heterogeneous imaging lower bound is based on the correspondence between the heterogeneous index values of the reference mask and the aforementioned "heterogeneous imaging upper bound" and the aforementioned "heterogeneous imaging lower bound" respectively. The block corresponding to the color in the rainbow level covers the reference pixel of the reference mask. The reference pixel is imaged and the feature image is imaged. By using a visualized image, the physician can use Molybdenum County to have each of these reference masks according to the above-described imaging rules, that is, the tumor heterogeneity shown in Fig. 12B is paid.

需注意的是，在腫瘤異質化特《彡像化時，除了持續It should be noted that in the case of tumor heterogeneity, in addition to persistence

同時地觀察此腫瘤所具有的其他特徵。 綜上所述，藉由本發明所袒μ & _ i &Other features of this tumor were observed simultaneously. In summary, by the present invention, μ & _ i &

52 201126260 方法、腫瘤囊腫特徵的影像化方法、腫瘤鈣化特徵的量化 方法、腫瘤鈣化特徵的影像化方法、腫瘤迴音性特徵的量 化方法、腫瘤異質化特徵的量化方法及腫瘤異質化特徵的 影像化方法，醫師可於拿到一腫瘤之超音波灰階影像的同 時，一併得到腫瘤這些特徵的量化數據與影像化圖像，做 為判斷腫瘤之性質的依據，以大幅提昇藉由腫瘤之超音波 灰階影像判斷腫瘤性質之程序的準確率及可靠度，且減輕 醫師在判斷腫瘤性質時的負擔。 除此之外’本發明所提供之腫瘤邊緣特徵的量化方 法、腫瘤邊緣特徵的影像化方法、腫瘤囊腫特徵的量化方 法、腫瘤囊腫特徵的影像化方法、腫瘤鈣化特徵的量化方 法、腫瘤鈣化特徵的影像化方法、腫瘤迴音性特徵的量化 方法、腫瘤異質化特徵的量化方法及腫瘤異質化特徵的影 像化方法所分別具有的各執行步驟，可以電腦語言寫成以 便執行’而該寫成之軟體程式可以儲存於任何微處理單元 可以辨識、解讀之紀錄媒體’或包含有該紀錄媒體之物品 及裝置。其不限為任何形式’該物品 碟，、—片、隨機存取記憶體 熟悉此項技藝者所可使用之包含有該紀錄媒體之物品。 上述實施例僅係為了方便說明而舉例而已 主張之權利範圍自應以巾請專利範圍所述為準，而非= 於上述實施例。 而非僅限 【圖式簡單說明】 201126260 圖1 A係習知之藉由手寫輸入方式標示出腫瘤輪廓的超音 波景;J像。 圖丨B係習知之藉由snake演算法計算並標示出腫瘤輪廓的 超音波影像。 圖2係顯示一電腦系統之架構的示意圖。 圖3 A係一超音波灰階影像的示意圖。 圖3B係本發明第一實施例之腫瘤邊緣特徵之量化方法的 流程圖。 圖3C係本發明第一實施例之腫瘤邊緣特徵之量化方法之 步驟（A)所應用之腫瘤輪廓擷取方法的流程圖。 圖3「D則為應用此「腫瘤輪廓掘取方法」以操取「腫瘤輪廊」 與腫瘤輪廓環形區域」之包含一腫瘤之灰階影像圖。 圖3E係顯示本發明第一實施例之腫瘤邊緣特徵之量化方 v驟（D)所包含之子步驟（ο】）及子步驟（D2)的流程圖。 圖3F係顯示本發明[實施例之腫瘤邊緣特徵之量化方 法之步驟（E)所包含之子步驟（E1)的流程圖。 圖4八釭本發明第二實施例之腫瘤邊緣特徵之影像化方法 的流裎圖。 圖4B係顯示本發明第二實施例之腫瘤邊緣特徵之影像化 化方法之步驟（D)所包含之子步驟（D1)及子步驟（1)2)的流 程圖。 系” ·、員示本發明第二實施例之腫瘤邊緣特徵之參像化 化方法之步驟⑹所包含之子步驟(E〇的流程圖。’、 201126260 ί象之超音波灰階影像 圖5係一顯示腫瘤邊緣特徵影像化圖 的示意圖。 圖6 Α係一超音波灰階影像的示意圖。 _ 系本發明第三實施例之腫瘤囊腫特徵之量化方法的 流程圖。 圖6C係顯示本發明第i實施例之腫瘤囊腫特徵之量化方 法之步驟（D)所包含之子步驟（D1)的流程圖。 圖7A係本發明第四實施例之腫瘤囊腫特徵之影像化方法 的流程圖。 圖7B係一顯示腫瘤囊腫特徵影像化圖像之超音波灰階影 像的不意圖。 圖8A係一超音波灰階影像的示意圖。 圖8B係本發明第五實施例之腫瘤鈣化特徵之量化方法的 流程圖。 圖8C係顯示本發明第五實施例之腫瘤鈣化特徵之量化方 法之步驟（F)所包含之子步驟（F1)的流程圖。 圖9A係本發明第六實施例之腫瘤鈣化特徵之影像化方法 的流程圖。 圖9B A 顯示腫瘤辦化特徵影像化圖像之超音波灰階影 像的示意圖。 圖10A係一超音波灰階影像的示意圖。 圖10B係本發明第七實施例之腫瘤迴音性特徵之量化方法 的流程圖。 圖丨丨Α係一超音波灰階影像的示意圖。 201126260 圖丨1B係本發明第八實施例之腫瘤異質化特徵之量化方法 的流程圖。 圖丨1C係顯示本發明第八實施例之腫瘤異質化特徵之量化 方法之步驟（F)所包含之子步驟（F1)的流程圖。 圖12A係本發明第九實施例之腫瘤異質化特徵之影像化方 法的流程圖。 圖丨2 β係一顯示腫瘤異質化特徵影像化圖像之超音波灰階 影像的示意圖。 【主要元件符號說明】 23記憶體 2 6糸統程式 35線段 21顯示裝置 22處理器 24輪入裝置 25儲存裝置 31~33 轨跡 34、36 點 5652 201126260 Methods, imaging methods for tumor cyst features, methods for quantifying tumor calcification characteristics, imaging methods for tumor calcification features, methods for quantifying tumor echogenic features, methods for quantifying tumor heterogeneity characteristics, and imaging of tumor heterogeneity features In this way, the physician can obtain the ultrasound grayscale image of the tumor and obtain the quantitative data and the image of the tumor as a basis for judging the nature of the tumor, so as to greatly enhance the tumor by the super The accuracy and reliability of the procedure for determining the tumor properties of the sonic grayscale image, and reducing the burden on the physician in judging the nature of the tumor. In addition, the method for quantifying tumor edge features, the imaging method for tumor edge features, the method for quantifying tumor cyst features, the imaging method for tumor cyst features, the method for quantifying tumor calcification characteristics, and the calcification characteristics of tumors are provided by the present invention. The imaging method, the method for quantifying the echogenicity of the tumor, the method for quantifying the heterogeneity of the tumor, and the imaging method for the heterogeneous feature of the tumor, respectively, can be written in a computer language to execute the software program written It can be stored in any recording medium that can be recognized and interpreted by any microprocessor unit or an item and device containing the recording medium. It is not limited to any form. The article disc, sheet, random access memory is familiar to those skilled in the art and may include articles containing the recording medium. The above-described embodiments are merely exemplified for convenience of description, and the scope of the claims is based on the scope of the patent application, rather than the above embodiment. Rather than limited [Simplified illustration] 201126260 Figure 1 A is a supersonic view of the contour of a tumor by means of handwriting input; J image. Figure B is a conventional calculation of the ultrasound image of the tumor contour by the snake algorithm. Figure 2 is a schematic diagram showing the architecture of a computer system. Figure 3 is a schematic diagram of an ultrasonic grayscale image. Fig. 3B is a flow chart showing a method of quantifying tumor edge features of the first embodiment of the present invention. Fig. 3C is a flow chart showing the method of tumor contour extraction applied in the step (A) of the method for quantifying tumor edge features according to the first embodiment of the present invention. Figure 3 "D is a grayscale image of a tumor containing the "tumor contouring method" to capture the "tumor ridge" and the tumor contour ring area". Fig. 3E is a flow chart showing sub-steps (o) and sub-steps (D2) included in the quantizing method (D) of the tumor edge feature of the first embodiment of the present invention. Fig. 3F is a flow chart showing the substep (E1) included in the step (E) of the method for quantifying the tumor edge feature of the present invention. Figure 4 is a flow chart of the imaging method of the tumor edge feature of the second embodiment of the present invention. Fig. 4B is a flow chart showing substep (D1) and substep (1) 2) included in the step (D) of the image forming method for the tumor edge feature of the second embodiment of the present invention. The sub-steps included in the step (6) of the parameterization method for the tumor edge feature of the second embodiment of the present invention (the flow chart of E〇. ', 201126260 ί象的 ultrasonic grayscale image 5 A schematic diagram showing a visualization of a tumor edge feature. Fig. 6 is a schematic diagram of a sputum-acoustic grayscale image. _ A flowchart of a method for quantifying tumor cyst characteristics according to a third embodiment of the present invention. Fig. 6C shows the first aspect of the present invention. Figure 7A is a flow chart showing a method for imaging a tumor cyst feature according to a fourth embodiment of the present invention. Figure 7B is a flow chart of a sub-step (D1) included in the step (D) of the method for quantifying tumor cyst characteristics of the embodiment. A schematic diagram showing an ultrasonic grayscale image of a tumor cyst characteristic image. Fig. 8A is a schematic diagram of an ultrasonic grayscale image. Fig. 8B is a flow chart of a method for quantifying tumor calcification characteristics according to a fifth embodiment of the present invention. Fig. 8C is a flow chart showing a substep (F1) included in the step (F) of the method for quantifying tumor calcification characteristics according to the fifth embodiment of the present invention. Fig. 9A is a tumor calcium according to a sixth embodiment of the present invention. FIG. 9B is a schematic diagram showing an ultrasonic grayscale image of a tumorized feature image. FIG. 10A is a schematic diagram of an ultrasonic grayscale image. FIG. 10B is a seventh embodiment of the present invention. A flow chart of a method for quantifying the echogenicity of a tumor. Fig. 1 is a flow chart of a method for quantifying the heterogeneity of tumors according to an eighth embodiment of the present invention. Figure 1C is a flow chart showing a substep (F1) included in the step (F) of the method for quantifying tumor heterogeneity characteristics according to the eighth embodiment of the present invention. Fig. 12A is a diagram showing the heterogeneous characteristics of the tumor according to the ninth embodiment of the present invention. Flow chart of the imaging method. Fig. 2 β system shows a schematic diagram of the ultrasonic grayscale image of the tumor heterogeneous feature image. [Main component symbol description] 23 memory 2 6 system program 35 line segment 21 display device 22 processor 24 wheeling device 25 storage device 31~33 track 34, 36 point 56

Claims (1)

201126260 七、申請專利範圍： 1. -種腫瘤邊緣特徵的量化方法，係應用於一由複數 個像素點組合而成並至少顯示-殖瘤的灰階影像，包括下 列步驟： (A) 從該灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環 形區域，且該腫瘤輪廓係位於該腫瘤輪廓環形區域内； (B) 將該腫瘤輪廓重疊顯示於該灰階影像上，以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 掏取泫腫瘤輪靡環形區域的一重心點，定義二從該 重心點向外延伸並通過該腫瘤輪廓環形區域的剖面線，及 提供一位於該剖面線上並位於該腫瘤輪廓環形區域内之量 測線段； (D) 計算出位於該量測線段上之該等像素點所分別具 有之灰階移動變異值；以及 (E) 依據位於該量測線段上之每一該等像素點所分別 具有之灰階移動變異值，將位於該剖面線上之腫瘤邊緣特 徵量化。 2. 如申請專利範圍第1項所述之量化方法，其中該步 驟（D)包括一步驟（D1 )，依據位於該量測線段上之該等像素 點所分別具有之灰階移動變異值，計算出位於該量測線段 上之該等像素點所具有之灰階移動變異值的標準差。 3. 如申請專利範圍第2項所述之量化方法，其中該步 驟（D)於該步驟（D1)之後更包括一步驟（D2) ’依據位於該量 測線段上之該等像素點所分別具有之灰階移動變異值，計 201126260 异出位於該量測線段上之該等像素點所具有之灰階移動變 異值的平均值。 4·如申請專利範圍第3項所述之量化方法，其中該步 驟（E)包括一步驟（E 1 )，藉由位於該量測線段上之該等像素 點所具有之灰階移動變異值的標準差及灰階移動變異值的 平均值，定義出一灰階移動變異值閥值，以判別位於該剖 面線上之腫瘤邊緣特徵的模糊程度。 5.如申請專利範圍第1項所述之量化方法，於該步驟 (E)之後更包括一步驟（F)改變該剖面線之位置，以掃瞄該腫 瘤之全部邊緣而將該腫瘤之全部邊緣特徵量化。 6· —種腫瘤邊緣特徵的影像化方法，係應用於一由複 數個像素點組合而成並至少顯示一腫瘤的灰階影像包括 下列步驟： (A) 從該灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環 形區域，且該腫瘤輪廓係位於該腫瘤輪廓環形區域内； (B) 將該腫瘤輪廓重疊顯示於該灰階影像上，以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 擷取該腫瘤輪廓環形區域的一重心點，定義一從該 重心點向外延伸並通過該腫瘤輪廓環形區域的剖面線，及 提供一位於泫剖面線上並位於該腫瘤輪廓環形區域内之量 測線段； (D) 計算出位於該量測線段上之該等像素點所分別具 有之灰階移動變異值；以及 201126260 (E)依據位於該量測線段上之該等像素點所分別具有 之灰階移動變異值，定義出一邊緣成像上界及一邊緣成像 下界’且配合-衫虹色階將位於該剖面線上之腫瘤邊緣特 徵影像化。 7. 如申請專利範圍第6項所述之影像化方法，其中該 步驟（D)包括一步驟（D1)，依據位於該量測線段上之該等像 素點所分別具有之灰P皆移動變異值，計#出位於該量測線 •k上之该等像素點所具有之灰階移動變異值的標準差。 8. 如申請專利範圍第7項所述之影像化方法，其中該 步驟（D)於該步驟（Di)之後更包括一步驟（D2)’依據位於該 量測線段上之該等像素點所分別具有之灰階移動變異值， 計算出位於該量測線段上之該等像素點所具有之灰階移動 變異值的平均值。 9. 如申請專利範圍第8項所述之影像化方法，其中該 步驟（E)包括一步驟（E ^藉由位於該量測線段上之該等像 素點所具有之灰階移動變異值的標準差及灰階移動變異值 的平均值疋義出該邊緣成像上界及該邊緣成像下界。 10. 如申請專利範圍第9項所述之影像化方法，其中該 邊緣成像上界係為該灰階移動變異值的平均值加上三倍之 該灰階移冑變異值的標準差，該邊緣成像下界則為該灰階 移動變異值的平均值減去三倍之該灰階移動變異值的標準 差。 1 1 ·如申请專利範圍第6項所述之影像化方法，其中該 衫虹色h ίτ、為一紅橙黃綠藍靛紫之連續漸變色階，且當一 59 201126260 或等於：1又上之像素點所具有之灰階移動變異值大於 化時便广成像上界時，該像素點在腫瘤邊緣特徵影像 化時便破一紅色區塊覆蓋。 ,2·如申*專利㈣6項所述之影像化方法，盆中該 彩虹色階係為-時黃綠⑼紫之連續漸變色階且當一 X里測線奴上之像素點所具有之灰階移動變異值小於 或等於該邊緣成像下界時’該像素點在腫瘤邊緣特徵影像 化時便被一紫色區塊覆蓋。 13. 如申明專利範圍第6項所述之影像化方法，其中該 彩虹色階係為-紅燈黃綠聽紫之連續漸變色階，^當^ 位於該量測線段上之像素點戶斤具有之灰階移動變異值介於 該邊緣成像上界及該邊緣成像下界之間時，該像素點在腫 瘤邊緣特徵影像化時便依據其灰階移動變異值分別與該邊 緣成像上界及該邊緣成像下界之間的對應關係，被一具有 從該彩虹色階中對應出之顏色的區塊覆蓋。 14. 如申請專利範圍第6項所述之影像化方法，於該步 驟（E)之後更包括一步驟（F)改變該剖面線之位置，以掃瞄該 腫瘤之全部邊緣而將該腫瘤之全部邊緣特徵影像化。 1 5. —種腫瘤囊腫特徵的量化方法，係應用於一由複數 個像素點組合而成並至少顯示一腫瘤的灰階影像，包括下 列步驟： (A)從該灰階影像操取出一腫瘤輪廓及一腫瘤輪廊環 形區域，且該腫瘤輪廓係位於該腫瘤輪廓環形區域内； 60 201126260 (B)將該腫瘤輪廓重疊顯示於該灰階影像上，以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (c)藉由位於該腫瘤内部區域内之該等像素點所分別 具有之灰階梯度值’計算出位於該膣瘤内部區域内之該等 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差；以及 (D)依據位於該腫瘤内部區域内之該等像素點所具有 之灰階梯度值的最小值及灰階梯度值的標準差，將位於該 腫瘤内部區域内之囊腫特徵量化。 1 6.如申請專利範圍第丨5項所述之量化方法，其中該步 驟（D)包括一步驟（D丨），藉由位於該腫瘤内部區域内之該等 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差’定義出一囊臛特徵之灰階梯度值的闊值，以計算出該 囊腫特徵於該腫瘤内部區域内所佔的比例。 1 7. —種腫瘤囊腫特徵的影像化方法，係應用於一由複 數個像素點組合而成並至少顯示一腫瘤的灰階影像，包括 下列步驟： (A) 從该灰階影像操取出一腫瘤輪廓及一腫瘤輪廓環 形區域，且該腫瘤輪廓係位於該腫瘤輪廓環形區域内； (B) 將該腫瘤輪廓重疊顯示於該灰階影像上，以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 藉由位於該腫瘤内部區域内之該等像素點所分別 具有之灰階梯度值，計算出位於該腫瘤内部區域内之該等 201126260 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差；以及 (D) 依據位於該腫瘤内部區域内之該等像素點所具有 之灰階梯度值的最小值及灰階梯度值的標準差，定義出一 囊腫成像上界及一囊腫成像下界，以將位於該腫瘤内部區 域内之腫瘤囊腫特徵影像化。 1 8.如申請專利範圍第丨7項所述之影像化方法，其中該 囊腫成像上界係為位於該腫瘤内部區域内之該等像素點所 具有之灰階梯度值的最小值加上零點一倍之該灰階梯度值 的標準差’該囊腫成像下界則為位於該腫瘤内部區域内之 該等像素點所具有之灰階梯度值的最小值。 19. 如申請專利範圍第丨7項所述之影像化方法，其中當 一位於該腫瘤内部區域内之該等像素點所具有之灰階梯度 值介於該囊腫成像上界及該囊腫成像下界之間時，該像素 點在腫瘤囊腫特徵影像化時便被一桃紅色區塊覆蓋。 20. 如申請專利範圍第丨7項所述之影像化方法，其中在 該步驟（D)後更包括下列步驟： (E) 將位於該腫瘤内部區域内之該等像素點分別定義 為複數個參考料，且每__該等參考料係包含—基準像 素點與複數個相鄰於該基準像素點之像素點且該基準像 素點所具有之灰梯度值係介於該囊腫成像上界及該囊腫 成像下界之間；以及 (F) 當至少一該等像素點所具有之灰階梯度值介於該囊 腫成像上界及該囊腫成像下界之間時，該基準像素點及該 201126260 等像素點在腫瘤囊腫特徵影像化時便被一桃紅色區塊覆 蓋。 2 1.如申請專利範圍第丨7項所述之影像化方法，其中在 該步驟（F)後更包括—步驟⑴），當只有該基準像素點所具有 之灰階梯度值介於該囊腫成像上界及該囊腫成像下界之間 時’移除覆蓋於該基準像素點及該等像素點的該桃紅色區 塊。 22. —種腫瘤鈣化特徵的量化方法，係應用於一由複數 個像素點組合而成並至少顯示一腫瘤的灰階影像，包括下 列步驟： (A) 從該灰階影像梅取出一腫瘤輪廓及一胺瘤輪廓環 形區域，且該腫瘤輪廓係位於該腫瘤輪廓環形區域内； (B) 將該腫瘤輪廓重疊顯示於該灰階影像上，以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 藉由位於該腫瘤内部區域内之該等像素點所分別 具有之灰階梯度值，計算出位於該腫瘤内部區域内之該等 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差； (D) 依據位於該腫瘤内部區域内之該等像素點所具有 之灰階梯度值的最小值及灰階梯度值的標準差，從該灰階 影像中擷取出一位於該腫瘤内部區域内之囊腫區域； (E) 藉由位於該腫瘤内部區域内但位於該囊腫區域之 外之該等像素點所分別具有之灰階梯度值，計算出位於該 腫瘤内部區域内但位於該囊腫區域之外之該等像辛點所具 63 201126260 有之灰階梯度值的最大值、灰階梯度值的標準差及灰階梯 度值的平均值；以及 (F)依據位於該腫瘤内部區域内但位於該囊腫區域之外 之該等像素點所具有之灰階梯度值的最大值、灰階梯度值 的標準差及灰階梯度值的平均值，將位於該腫瘤内部區域 内之腫瘤鈣化特徵量化。 23. 如申請專利範圍第22項所述之量化方法，其令該步 驟（F)包括一步驟（F 1 )，藉由位於該腫瘤内部區域内但位於 該囊腫區域之外之該等像素點所具有之灰階梯度值的最大 值、灰階梯度值的標準差及灰階梯度值的平均值，定義出 一鈣化特徵之灰階梯度值的閥值，以計算出該鈣化特徵於 該腫瘤内部區域内所佔的比例。 24. —種腫瘤鈣化特徵的影像化方法，係應用於一由複 數個像素點組合而成並至少顯示一腫瘤的灰階影像，包括 下列步驟： (A) 從該灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環 形區域，且該腫瘤輪廓係位於該腫瘤輪廓環形區域内； (B) 將該腫瘤輪廓重疊顯示於該灰階影像上，以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 藉由位於該腫瘤内部區域内之該等像素點所分別 具有之灰階梯度值’計算出位於該腫瘤内部區域内之該等 像素點所具有之灰階梯度值的最小值及灰階梯度值的標準 差； 64 201126260 (〇)依據位於該腫瘤内部區域内之該等像素點所具有 之灰階梯度值的最小值及灰階梯度值的標準差，從該灰階 影像中擷取出一位於該腫瘤内部區域内之囊腫區域； (E) 藉由位於該腫瘤内部區域内但位於該囊腫區域之 外之該等像素點所分別具有之灰階梯度值，計算出位於該 腫瘤内部區域内但位於該囊腫區域之外之該等像素點所具 有之灰階梯度值的最大值、灰階梯度值的標準差及灰階梯 度值的平均值；以及 (F) 依據位於該腫瘤内部區域内但位於該囊腫區域之外 之該等像素點所真有之灰階梯度值的最大值、灰階梯度值 的標準差及灰階梯度值的平均值，定義出一鈣化成像上界 及一鈣化成像下界’以將位於該腫瘤内部區域内之腫瘤鈣 化特徵影像化。 25. 如申請專利範圍第24項所述之影像化方法，其中該 名弓化成像上界係為位於該腫瘤内部區域内但位於該囊腫區 域之外之該等像素點所具有之灰階梯度值的最大值，該鈣 化成像下界則為位於該腫瘤内部區域内但位於該囊腫區域 之外之該等像素點所具有之灰階梯度值的平均值加上二點 八倍之該灰階梯度值的標準差。 26. 如申請專利範圍第24項所述之影像化方法，其中當 一位於該腫瘤内部區域内但位於該囊腫區域之外之該等像 素點所具有之灰階梯度值介於該鈣化成像上界及該鈣化成 像下界之間時，該像素點在腫瘤鈣化特徵影像化時便被一 黃色區塊覆蓋。 201126260 27. —種腫瘤迴音性特徵的量化方法，係應用於一由複 數個像素點組合而成並至少顯示__腫瘤的灰階影像包括 下列步驟： (A) 應從該灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓 %形區域，且該腫瘤輪廓係位於該腫瘤輪廓環形區域内； (B) 將泫腫瘤輪廓重疊顯示於該灰階影像上以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 藉由位於該腫瘤内部區域内之該等像素點所分別 具有之灰階梯度值，計算出位於該腫瘤内部區域内之該等 像素點所具有之灰階梯度值的平均值； (D) 於該腫瘤外部區域選取一基準區域，藉由位於該基 準區域内之該等像素點所分別具有之灰階梯度值，計算出 位於該基準區域内之該等像素點所具有之灰階梯度值的平 均值；以及 (E) 依據位於該腫瘤内部區域内之該等像素點所具有 之灰階梯度值的平均值及位於該基準區域内之該等像素點 所具有之灰階梯度值的平均值，將該腫瘤之迴音性特徵量 化。 28. 如申請專利範圍第27項所述之量化方法，其中該迴 音性特徵係藉由將位於該腫瘤内部區域内之該等像素點所 具有之灰階梯度值的平均值減去位於該基準區域内之該等 像素點所具有之灰階梯度值的平均值所得到之差值除以位 於該基準區域内之該等像素點所具有之灰階梯度值的平均 值的方式被量化。 66 201126260 29.如申請專利範圍第27項所述之量化方法其中當迴 音^特徵量化所得出的數值大於或等於零時，該腫瘤便具 有高迴音性之特徵。 立3〇,如申請專利範圍第27項所述之量化方法，其中當迴 曰1·生特徵I化所得出的數值小於零時，該腫瘤便具有低迴 音性之特徵。 31'種腫瘤異質化特徵的量化方法，係應用於一由複 數個像素點組合而成並至少顯示—腫瘤的灰階影像包括 下列步驟： 。（A)k忒灰階影像擷取出一腫瘤輪廓及一腫瘤輪廓環 形區域，且泫腫瘤輪廓係位於該腫瘤輪廓環形區域内； (B) 將该腫瘤輪廓重疊顯示於該灰階影像上，以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域，. (C) 將位於該腫瘤内部區域内之該等像素點分別定義 為複數個參考遮| ’且每—該等參考遮軍係包含—基準像 素點與複數個相鄰於該基準像素點之像素點·， (D) 計算出每—該等參考遮罩所分別具有之參考遮罩 灰階梯度值局部平均及參考㈣灰階梯度值局部變異； (E) 計算出每-該等參考遮罩所分別具有之參考遮罩 灰階梯度值局部平均之變異、參考遮罩灰階梯度值局部變 異之平均以及參考遮罩灰階梯度值局部變異之變異；以及 (F) 藉由每-該等參考遮罩所分別具有之至少―選自於 -由參考遮罩灰階梯度值局部平均之變異、參考遮罩灰階 梯度值局部變異之平均以及參考遮罩灰階梯度值局 201126260 之變異所構成之群組，計算出每一該等參考遮罩所分別具 有之異質化指標值，將該腫瘤之異質化特徵量化。 3 2 _如申請專利範圍第3 1項所述之量化方法，其中每一 該等參考遮罩係包含25個像素點 3 3 ·如申明專利範圍第3 1項所述之量化方法’其中步驟 (F)包含一步驟（Fi)，藉由辟該參考遮罩所具有之參考遮罩 灰階梯度值變異減去該等參考遮罩所具有之平均遮罩灰階 梯度值變異所得之差值的絕對值開根號的方式，計算出每 一該等參考遮罩所分別具有之異質化指標值。 34. 如申請專利範圍第3 1項所述之量化方法，其中該異 質化指標值係一變數為參考遮罩灰階梯度值局部平均之變 異的函數。 35. 如申請專利範圍第3丨項所述之量化方法其中該異 質化指標值係一變數為參考遮罩灰階梯度值局部變異之平 均的函數。 36. 如申請專利範圍第3 1項所述之量化方法其中該異 質化指標值係一變數為參考遮罩灰階梯度值局部變異之變 異的函數。 37·如申請專利範圍第3 1項所述之量化方法，其中該異 貝化指標值係一變數為參考遮罩灰階梯度值局部變異之變 異除以參考遮罩灰階梯度值局部平均之變異所得之比值的 函數。 3 8 ·如申請專利範圍第3 1項所述之量化方法，其中該異 貝化指標值係一變數為參考遮罩灰階梯度值局部變異之平 68 201126260 均除以參考遮罩灰階梯度值局部變異之變異所得之比值的 函數。 39.如申請專利範圍第3丨項所述之量化方法，其中該異 質化指標值係一變數為參考遮罩灰階梯度值局部變異之平 均除以參考遮罩灰階梯度值局部平均之變異所得之比值的 函數。 40. —種腫瘤異質化特徵的影像化方法，係應用於一由 複數個像素點組合而成並至少顯示一腫瘤的灰階影像，包 括下列步驟： (A) 從泫灰階影像擷取出一腫瘤輪廊及一腫瘤輪廓環 形區域’且該腫瘤輪廓係位於該腫瘤輪廓環形區域内； (B) 將該腫瘤輪廓重疊顯示於該灰階影像上以在該灰 階影像上定義出一腫瘤内部區域及一腫瘤外部區域； (C) 將位於該腫瘤内部區域内之該等像素點分別定義 為複數個參考遮罩’且每—該等參考遮罩係包含—基準像 素點與複數個相鄰於該基準像素點之像素點； (D) 藉由包含於每一該等參考遮罩之該等像素點所分 別具有之灰階梯度值’計算出每—該等參考遮罩所分別具 有之參考遮罩灰階梯度值變異； ⑹藉由每—該等參考遮罩所分別具有之參考遮罩灰 階梯度值變異，計算出該等參考遮罩所具有之平均遮罩灰 階梯度值變異； (F)藉由每一該等參考遮革所分別具有之參考遮罩灰階 梯度值變異及該等參考遮罩 1、皁所具有之平均遮罩灰階梯度值 201126260 變異，計算出每一該等參考遮革所分別具有之異質化指標 值； (G) “由每一泫等參考遮罩所分別具有之異質化指標 值，計鼻出戌等參考遮罩所具有之異質化指標值的最大 值、異質化指標值的最小值、異質化指標值的平均值及異 質化指標值的標準差；以及 (H) 依據該等參考遮罩所具有之異質化指標值的最大 值、異質化指標值的最小值、異質化指標值的平均值及異 質化指標值的標準差，〖義出一異質化成像上界及一異質 化成像下界，且配合—彩虹色階將位於該腫瘤内部區域内 之腫瘤異質化特徵影像化。 4 1 ·如申請專利範圍第4〇項所述之影像化方法其中該 異質化成像上界係為該等參考遮罩所具有之異質化指標值 的最大值，該異質化成像下界則為該等參考遮罩所具有之 異質化指標值的平均值減去零點__倍之異質化指標值的標 準差。 ' 42.如申請專利範圍第4〇項所述之影像化方法，其中該 彩虹色階係為一紅橙黃綠藍靛紫之連續漸變色階，且當— 參考遮罩所具有之異質化指標值等於該異質化成像上界 時，忒參考遮罩之基準像素點在腫瘤異質化特徵影像化時 便被一紅色區塊覆蓋。 43·如申請專利範圍第4〇項所述之影像化方法，其中該 心虹色係為一紅橙黃綠藍鼓紫之連續漸變色階且當— >考遮罩所具有之異質化指標值小於或等於該異質化成像 70 201126260 下界時，袭_ Jy Μ，号遮罩之基準像素點在腫瘤異質化特徵影 化時便被一紫色區塊覆蓋。 '44.如申凊專利範圍第項所述之影像化方法，其中該 杉虹色階係為—紅橙黃綠藍靛紫之連續漸變色階且當一 參考遮罩所具有之異質化指標值介於該異質化成像上界及 該異$化成像下界之間時，該參考遮罩之基準像素點在腫 瘤異貝化特徵影像化時便依據其異質化指標值分別與該異 質化成像上界及該異質化成像下界之間的對應關係，被一 具有彳之该彩虹色階中對應出之顏色的區塊覆蓋。 45·如申請專利範圍第4〇項所述之影像化方法其中每 一該等參考遮罩係包含25個像素點。201126260 VII. Patent application scope: 1. A method for quantifying tumor edge features, which is applied to a grayscale image composed of a plurality of pixels and displaying at least a tumor, including the following steps: (A) from the The grayscale image extracts a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the grayscale image to be in the grayscale image Defining an inner region of the tumor and an outer region of the tumor; (C) drawing a center of gravity of the annular region of the tumor rim, defining a section line extending outward from the center of gravity and passing through the annular region of the contour of the tumor, and Providing a measurement line segment located on the section line and located in the annular region of the tumor contour; (D) calculating gray scale movement variation values respectively of the pixels located on the measurement line segment; and (E) basis Each of the pixels located on the measurement line has a gray-scale movement variation value, and the tumor edge features located on the section line are quantized. 2. The quantization method according to claim 1, wherein the step (D) comprises a step (D1) according to a gray-scale movement variation value of the pixels respectively located on the measurement line segment, Calculating the standard deviation of the gray-scale movement variation values of the pixels located on the measurement line segment. 3. The method of claim 2, wherein the step (D) further comprises a step (D2) after the step (D1), wherein the pixels are located on the measurement line segment. With the gray-scale moving variation value, the average value of the gray-scale moving variation value of the pixels located on the measuring line segment is different from 201126260. 4. The method of claim 3, wherein the step (E) comprises a step (E1), wherein the gray-scale movement variation value of the pixels located on the measurement line segment has The standard deviation and the average of the gray-scale movement variograms define a gray-scale moving variability threshold to determine the degree of blurring of the tumor edge features located on the section line. 5. The quantification method according to claim 1, wherein after step (E), a step (F) is further included to change the position of the hatching to scan all the edges of the tumor and the tumor is all Edge feature quantization. 6. The imaging method for the edge feature of a tumor is applied to a grayscale image composed of a plurality of pixels and displaying at least one tumor, comprising the following steps: (A) extracting a tumor contour from the grayscale image And a tumor contour annular region, wherein the tumor contour is located in the annular contour region of the tumor; (B) displaying the tumor contour on the gray scale image to define an internal tumor region and the grayscale image An outer region of the tumor; (C) drawing a center of gravity of the annular region of the contour of the tumor, defining a section line extending outward from the point of gravity and passing through the annular region of the contour of the tumor, and providing a line on the ridge and located at a measurement line segment in the annular region of the tumor contour; (D) calculating a gray scale movement variation value respectively for the pixels located on the measurement line segment; and 201126260 (E) depending on the measurement line segment The gray-scale moving variation values of the pixels respectively define an edge imaging upper bound and an edge imaging lower bound' and the matching-shirt rainbow level will be located on the section line. The tumor edge features are imaged. 7. The imaging method according to claim 6, wherein the step (D) comprises a step (D1), wherein the gray pixels P respectively move according to the pixels located on the measuring line segment The value, the standard deviation of the gray-scale movement variation values of the pixels located on the measurement line. 8. The imaging method according to claim 7, wherein the step (D) further comprises a step (D2) after the step (Di), according to the pixels located on the measuring line segment. The gray-scale moving variation values are respectively obtained, and the average value of the gray-scale moving variation values of the pixels located on the measuring line segment is calculated. 9. The imaging method of claim 8, wherein the step (E) comprises a step (E^ by the gray-scale movement variation value of the pixels located on the measurement line segment) The mean value of the standard deviation and the gray-scale moving variation value is the upper boundary of the edge imaging and the lower boundary of the edge imaging. 10. The imaging method according to claim 9, wherein the edge imaging upper bound is The average value of the gray-scale moving variation value plus three times the standard deviation of the gray-scale moving variation value, and the edge imaging lower bound is the average value of the gray-scale moving variation value minus three times the gray-scale moving variation value Standard deviation. 1 1 · The imaging method described in claim 6 wherein the shirt is a rainbow color h ίτ, which is a continuous gradient of red orange yellow green blue purple purple, and when a 59 201126260 or equals: When the pixel position of the pixel is larger than the upper limit of the image, the pixel will be covered by a red block when the edge feature of the tumor is imaged. 2, such as Shen* patent (4) 6 Imaging method described in the item, basin The rainbow color system is a continuous gradation of the yellow-green (9) purple color, and when the pixel of the X-ray line slave has a gray-scale moving variation value less than or equal to the lower boundary of the edge imaging, the pixel is in the tumor The edge feature is covered by a purple block when it is imaged. 13. The image method according to claim 6, wherein the rainbow color system is a continuous gradient color tone of red light yellow green listening purple, ^ When the gray-scale moving variation value of the pixel point on the measuring line segment is between the upper edge of the edge imaging and the lower boundary of the edge imaging, the pixel point is based on the gray when the tumor edge feature is imaged. The correspondence between the step movement variability value and the edge imaging upper bound and the edge imaging lower bound, respectively, is covered by a block having a color corresponding to the rainbow gradation. 14. As claimed in claim 6 In the imaging method, after the step (E), a step (F) is further included to change the position of the hatching to scan all the edges of the tumor to visualize all the edge features of the tumor. Tumor cyst The quantification method is applied to a gray scale image composed of a plurality of pixel points and displaying at least one tumor, comprising the following steps: (A) manipulating a tumor contour and a tumor wheel corridor ring from the gray scale image a region, and the tumor contour is located in the annular region of the tumor contour; 60 201126260 (B) overlaying the tumor contour on the gray scale image to define an internal tumor region and an external tumor on the gray scale image a region; (c) calculating a minimum grayscale value of the pixels located in the interior region of the tumor by the gray gradient value of each of the pixels located in the interior region of the tumor The standard deviation of the value and the gray gradient value; and (D) the minimum value of the gray gradient value and the standard deviation of the gray gradient value according to the pixels located in the inner region of the tumor, which will be located inside the tumor Quantification of cyst characteristics within the area. 1 6. The method of quantification according to item 5 of the patent application, wherein the step (D) comprises a step (D丨), wherein the gray points of the pixels located in the inner region of the tumor have The minimum value of the value and the standard deviation of the gray gradient value define a threshold value of the gray gradient value of a capsule characteristic to calculate the proportion of the cyst characteristic in the inner region of the tumor. 1 7. An imaging method for the characteristics of a tumor cyst is applied to a gray scale image composed of a plurality of pixels and displaying at least one tumor, comprising the following steps: (A) operating a grayscale image from the grayscale image a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the grayscale image to define a tumor interior on the grayscale image a region and an outer region of the tumor; (C) calculating, by the gray gradient values of the pixels located in the inner region of the tumor, the grays of the 201126260 pixels located in the inner region of the tumor a minimum value of the step value and a standard deviation of the gray step value; and (D) a minimum value of the gray step value and a standard deviation of the gray step value according to the pixels located in the inner region of the tumor, A cyst imaging upper bound and a cyst imaging lower bound are defined to visualize tumor cyst features located within the tumor's internal region. 1 8. The imaging method according to claim 7, wherein the upper boundary of the cyst imaging is a minimum value of a gray gradient value of the pixels located in an inner region of the tumor plus zero The standard deviation of the gray gradient value is doubled. The lower boundary of the cyst imaging is the minimum value of the gray gradient value of the pixels located in the inner region of the tumor. 19. The imaging method according to claim 7, wherein the pixels in the inner region of the tumor have a gray gradient value between the upper boundary of the cyst imaging and the lower boundary of the cyst imaging. When in between, the pixel is covered by a pink block when the tumor cyst features are imaged. 20. The imaging method according to claim 7, wherein the step (D) further comprises the following steps: (E) defining the pixels located in the inner region of the tumor as a plurality of pixels respectively. a reference material, and each of the reference materials includes a reference pixel point and a plurality of pixel points adjacent to the reference pixel point and the gray gradient value of the reference pixel point is between the upper boundary of the cyst image and The cyst is imaged between the lower bounds; and (F) when at least one of the pixels has a gray gradient value between the upper boundary of the cyst imaging and the lower boundary of the cyst imaging, the reference pixel and the 201126260 pixel The point is covered by a pink block when the tumor cyst features are imaged. 2 1. The imaging method according to claim 7, wherein after the step (F), the method further comprises: step (1), when only the reference pixel has a gray gradient value between the cyst The pink upper block covering the reference pixel point and the pixels is removed when the upper boundary of the image and the lower boundary of the cyst image are imaged. 22. A method for quantifying tumor calcification characteristics, which is applied to a gray scale image composed of a plurality of pixel points and displaying at least one tumor, comprising the following steps: (A) taking a tumor contour from the gray scale image plum And an amine tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the grayscale image to define an internal tumor region on the grayscale image And an outer region of the tumor; (C) calculating the gray gradient of the pixels located in the inner region of the tumor by the gray gradient values respectively obtained by the pixels located in the inner region of the tumor The minimum value of the value and the standard deviation of the gray gradient value; (D) from the gray value based on the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor a cystic region located in the inner region of the tumor is extracted from the image; (E) each of the pixels located in the inner region of the tumor but outside the region of the cyst The step value is calculated as the maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the gray gradient of the gray point value of the 2011 201126260 which is located in the inner region of the tumor but outside the cyst region. The average value of the values; and (F) the maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the gray gradient value of the pixels according to the pixels located in the inner region of the tumor but outside the cyst region. The average of the values quantifies the calcification characteristics of the tumor located within the interior region of the tumor. 23. The method of quantification according to claim 22, wherein the step (F) comprises a step (F1) by the pixels located in the inner region of the tumor but outside the cyst region The maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the average value of the gray gradient value define a threshold value of the gray gradient value of the calcification characteristic to calculate the calcification characteristic of the tumor The proportion of the internal area. 24. An imaging method for tumor calcification characteristics, which is applied to a gray scale image composed of a plurality of pixel points and displaying at least one tumor, comprising the following steps: (A) extracting a tumor from the gray scale image a contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) displaying the tumor contour on the grayscale image to define an internal tumor region on the grayscale image And an outer region of the tumor; (C) calculating the gray gradient of the pixels located in the inner region of the tumor by the gray gradient value of the pixels located in the inner region of the tumor The minimum value of the value and the standard deviation of the gray gradient value; 64 201126260 (〇) based on the minimum value of the gray gradient value and the standard deviation of the gray gradient value of the pixels located in the inner region of the tumor The grayscale image extracts a cystic region located within the interior region of the tumor; (E) the pixels located within the interior region of the tumor but outside the cystic region Having a gray gradient value, respectively, calculating the maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the gray gradient of the pixels located in the inner region of the tumor but outside the cyst region The average value of the values; and (F) the maximum value of the gray gradient value, the standard deviation of the gray gradient value, and the gray gradient according to the pixels located in the inner region of the tumor but outside the cyst region. The mean value of the values defines a calcified imaging upper bound and a calcified imaging lower bound' to visualize tumor calcification features located within the tumor's internal region. 25. The imaging method of claim 24, wherein the arched imaging upper boundary is a gray gradient of the pixels located in the inner region of the tumor but outside the cyst region. The maximum value of the calcification lower bound is the average value of the gray gradient values of the pixels located in the inner region of the tumor but outside the cyst region plus two to eight times the gray gradient The standard deviation of the values. 26. The imaging method of claim 24, wherein a gray level value of the pixels located in the inner region of the tumor but outside the cyst region is between the calcification imaging When the boundary between the boundary and the calcification image is between the boundaries, the pixel is covered by a yellow block when the tumor calcification feature is imaged. 201126260 27. A method for quantifying tumor echogenicity is applied to a grayscale image composed of a plurality of pixel points and displaying at least a __tumor comprising the following steps: (A) A grayscale image should be taken out of the grayscale image a tumor contour and a tumor contour %-shaped region, and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the gray scale image to define a tumor interior on the gray scale image a region and an outer region of the tumor; (C) calculating, by the gray gradient values of the pixels located in the inner region of the tumor, the gray steps of the pixels located in the inner region of the tumor (D) selecting a reference region in the outer region of the tumor, and calculating the grayscale value values of the pixels located in the reference region, respectively, in the reference region The average value of the gray gradient values of the pixels; and (E) the average value and location of the gray gradient values of the pixels located in the inner region of the tumor The average value of the gray gradient values of the pixels in the reference region quantifies the echogenic characteristics of the tumor. 28. The method of quantizing according to claim 27, wherein the echogenic feature is obtained by subtracting an average value of gray scale values of the pixels located in the inner region of the tumor from the reference The difference between the average value of the gray gradient values of the pixels in the region is quantized by dividing the average of the gray gradient values of the pixels located in the reference region. The method of claim 27, wherein the tumor has a high echogenicity when the value obtained by quantizing the echo feature is greater than or equal to zero. According to the quantification method described in claim 27, wherein the tumor has a low echogenicity when the value obtained by the trait 1 is less than zero. A method for quantifying 31' tumor heterogeneity features is applied to a combination of a plurality of pixel points and at least displayed - the grayscale image of the tumor includes the following steps: (A) k忒 grayscale image extracting a tumor contour and a tumor contour annular region, and the tumor contour is located in the annular contour region of the tumor; (B) superimposing the tumor contour on the grayscale image to An inner region of the tumor and an outer region of the tumor are defined on the grayscale image. (C) the pixels located in the inner region of the tumor are respectively defined as a plurality of reference masks |' and each of the reference masks The military system includes a reference pixel point and a plurality of pixel points adjacent to the reference pixel point, (D) calculating a reference mask gray step value local average and reference for each of the reference masks (4) Gray gradient value local variation; (E) Calculate the local average variation of the reference mask gray gradient value for each of the reference masks, the average of the reference mask gray gradient value local variation, and the reference mask a variation of the local variation of the gray gradient value; and (F) by each of the reference masks having at least "selected from - a local average variation of the reference mask gray scale value, a reference mask gray ladder degree Average local variation of the gray scale mask and the group reference value Bureau variation gradient composed of 201 126 260, the index value is calculated for each of these heterostructure reference masks having respectively, the heterogeneity of the tumor characteristics of the quantization. 3 2 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ (F) includes a step (Fi) by subtracting the variation of the average mask gray gradient value of the reference mask from the reference mask gray gradient value variation of the reference mask The absolute value of the root opening number is used to calculate the heterogeneity index value of each of the reference masks. 34. The method of quantification of claim 31, wherein the heterogeneous index value is a function of a variation of a local average of a reference gray scale value. 35. The quantification method of claim 3, wherein the heterogeneity index value is a function of an average of the local variation of the reference mask gray gradient value. 36. The quantification method of claim 31, wherein the heterogeneous index value is a function of a variation of a local variation of a reference gray scale value. 37. The quantification method according to claim 31, wherein the heterobeat index value is a variation of the reference mask gray gradient value local variation variation divided by the reference mask gray gradient value local average A function of the ratio of the variance. 3 8 · The quantification method according to claim 31, wherein the heterobeat index value is a variable for the reference mask gray gradient value local variation flat 68 201126260 is divided by the reference mask gray gradient A function of the ratio of the variation of the local variation. 39. The method of claim 3, wherein the heterogeneous index value is a variable that is the average of the local variability of the reference mask gray gradient value divided by the local average of the reference mask gray gradient value. A function of the resulting ratio. 40. An imaging method for tumor heterogeneity is applied to a grayscale image composed of a plurality of pixels and displaying at least one tumor, comprising the following steps: (A) taking one from the grayscale image a tumor rim and a tumor contour annular region' and the tumor contour is located in the annular contour region of the tumor; (B) overlaying the tumor contour on the grayscale image to define a tumor interior on the grayscale image a region and an outer region of the tumor; (C) defining the pixels located within the region of the tumor as a plurality of reference masks, respectively, and each of the reference masks includes a reference pixel point and a plurality of adjacent a pixel point of the reference pixel; (D) calculating, by each of the pixels included in each of the reference masks, a gray gradient value of each of the reference masks Reference mask gray gradient value variation; (6) calculating the average mask gray gradient value of the reference mask by using the reference mask gray gradient value variation of each of the reference masks respectively (F) calculated by the reference mask gray gradient value variation of each of the reference shades and the average mask gray gradient value 201126260 of the reference mask 1 and the soap The heterogeneity index value of each of the reference occlusions; (G) “The heterogeneity index value of each reference mask, and the heterogeneity of the reference mask such as the nose sputum The maximum value of the index value, the minimum value of the heterogeneity index value, the average value of the heterogeneity index value, and the standard deviation of the heterogeneity index value; and (H) the maximum value of the heterogeneous index value according to the reference masks The minimum value of the heterogeneity index value, the average value of the heterogeneity index value, and the standard deviation of the heterogeneity index value, the upper bound of the heterogeneous imaging and the lower bound of a heterogeneous imaging, and the matching - the rainbow color gradation will be located Imaging heterogeneity of tumors in the inner region of the tumor. 4 1 · The imaging method described in claim 4, wherein the heterogeneous imaging upper bound is the heterogeneous index value of the reference masks The maximum value of the heterogeneous imaging is the standard deviation of the heterogeneous index values of the reference masks minus the zero-point __ times the standard deviation of the heterogeneity index values. ' 42. The imaging method according to the item, wherein the rainbow color system is a continuous gradient color gradation of red orange, yellow, green, blue, purple, and when the reference mask has a heterogeneity index value equal to the upper bound of the heterogeneous image, The reference pixel of the reference mask is covered by a red block when the tumor heterogeneous feature is imaged. 43. The imaging method according to claim 4, wherein the heart color is a red The orange-orange-blue-blue-violet continuous gradient color gradation and when the -> test mask has a heterogeneous index value less than or equal to the heterogeneous imaging 70 201126260 lower bound, hit _ Jy Μ, the reference pixel of the mask is at Tumor heterogeneity features are covered by a purple block when they are visualized. '44. The imaging method according to claim 1, wherein the cedar color system is a continuous gradation of red orange, yellow, green, blue, purple, and a heterogeneous index value of a reference mask. When the heterogeneous imaging upper bound and the lower bound of the different imaging image, the reference pixel of the reference mask is imaged according to the heterogeneous index value and the heterogeneous imaging upper bound respectively And the correspondence between the lower bounds of the heterogeneous imaging is covered by a block having a corresponding color in the rainbow gradation. 45. The imaging method of claim 4, wherein each of the reference masks comprises 25 pixels.
TW99101930A 2010-01-25 2010-01-25 Quantification method of the feature of a tumor and an imaging method of the same TWI474284B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW99101930A TWI474284B (en) 2010-01-25 2010-01-25 Quantification method of the feature of a tumor and an imaging method of the same

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW99101930A TWI474284B (en) 2010-01-25 2010-01-25 Quantification method of the feature of a tumor and an imaging method of the same

Publications (2)

Publication Number Publication Date
TW201126260A true TW201126260A (en) 2011-08-01
TWI474284B TWI474284B (en) 2015-02-21

Family

ID=45024456

Family Applications (1)

Application Number Title Priority Date Filing Date
TW99101930A TWI474284B (en) 2010-01-25 2010-01-25 Quantification method of the feature of a tumor and an imaging method of the same

Country Status (1)

Country Link
TW (1) TWI474284B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI569778B (en) * 2014-11-06 2017-02-11 安克生醫股份有限公司 Image enhancement method and imaging acceleration system for ultrasound scatterer signal

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5224175A (en) * 1987-12-07 1993-06-29 Gdp Technologies, Inc. Method for analyzing a body tissue ultrasound image
US6078680A (en) * 1997-07-25 2000-06-20 Arch Development Corporation Method, apparatus, and storage medium for detection of nodules in biological tissue using wavelet snakes to characterize features in radiographic images
US6173083B1 (en) * 1998-04-14 2001-01-09 General Electric Company Method and apparatus for analyzing image structures
ATE514144T1 (en) * 2001-10-16 2011-07-15 Univ Chicago COMPUTER-ASSISTED DETECTION OF THREE-DIMENSIONAL LESIONS

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI569778B (en) * 2014-11-06 2017-02-11 安克生醫股份有限公司 Image enhancement method and imaging acceleration system for ultrasound scatterer signal

Also Published As

Publication number Publication date
TWI474284B (en) 2015-02-21

Similar Documents

Publication Publication Date Title
Scalco et al. Texture analysis of medical images for radiotherapy applications
Lee et al. Automated segmentation of lesions including subretinal hyperreflective material in neovascular age-related macular degeneration
Han et al. A texture feature analysis for diagnosis of pulmonary nodules using LIDC-IDRI database
CN104363815B (en) Image processing apparatus and image processing method
CN101706843B (en) Interactive film Interpretation method of mammary gland CR image
Zhao et al. Saliency driven vasculature segmentation with infinite perimeter active contour model
US20100254584A1 (en) Automated method for assessment of tumor response to therapy with multi-parametric mri
CN110036408B (en) Automatic ct detection and visualization of active bleeding and blood extravasation
Moran et al. Using super-resolution generative adversarial network models and transfer learning to obtain high resolution digital periapical radiographs
JP5576775B2 (en) Image processing apparatus, image processing method, and image processing program
Chen et al. Semi-automatic geographic atrophy segmentation for SD-OCT images
CN103348359A (en) 3D retinal disruptions detection using optical coherence tomography
Odstrcilik et al. Thickness related textural properties of retinal nerve fiber layer in color fundus images
TWI624807B (en) Iterative analysis of medical images
WO2007026598A1 (en) Medical image processor and image processing method
JP5600426B2 (en) Nail plate pigment line discrimination device
Chen et al. Computerized detection and quantification of microcalcifications in thyroid nodules
RU2015130908A (en) AUTOMATED INTERESTING FABRIC IDENTIFICATION
JP2006346094A (en) Method for outputting detected information and medical image processing system
Habibollahi et al. Hyperechoic renal masses: differentiation of angiomyolipomas from renal cell carcinomas using tumor size and ultrasound radiomics
US20130236079A1 (en) Method and device for analysing a region of interest in an object using x-rays
TW201126260A (en) Quantification method of the feature of a tumor and an imaging method of the same
JP6519703B2 (en) Image processing method, diagnostic device, and program
US9773307B2 (en) Quantification and imaging methods and system of the echo texture feature
JP6254477B2 (en) Image processing apparatus, image processing method, and image processing program