1259765 五、發明說明（1) 【發明所屬之技 本發明利用 膜影像，以步進 格控制之白色冷 析關鍵區域之影 庠中，以達成臨 者 先韵 按 者，導 中及沉 或轉深 結石、 血症、 包括抽 、内視 來診斷 技術】 ，常見之 因於膽紅 積於組織 之現象。 膽管阻塞 先天性疾 血及腹部 鏡膽胰攝 術領域】 電腦影像處理技術，將臨床串、 :動之攝影鏡頭掘取患者眼； 光源環境下，以特有之辨色，二在嚴 像色調，並將分析結果匯入 S割及分 床診斷分級數據化及臨床細 t餐驗資料 、、工驗貧料庫累積 「黃疸」症狀常發於初生兒或肝膽疾 素及其成份黃色膽汁色素大量出現 = 間，引起皮膚、眼白及尿液三者顏色變ί 導致黃疸的原因有肝臟疾病、肝炎、膽囊 、溶血，或其他原因如：藥物副作用、Q敗 病…等。臨床上初步檢查黃疸的方法通常 超音波檢查，必要時還要進一步做肝穿刺 影、電腦斷層掃描及核磁共振掃描等處置 又東方人的膚色本來就是偏黃色，因此皮膚變黃時， '、t 4寸別注思’否則$被疏忽而不自知。通常眼球鞏膜變 汽早餐於病患皮膚顏色轉深，所以初期診斷之關鍵便在於 眼球拿膜之顏色是否轉黃。同時根據臨床醫師指出輕微黃 疫症狀表現是以眼白變黃接著皮膚轉黃；黃疸病狀消失時 也疋眼睛顏色優先於皮膚恢復正常。但醫師亦指出誤飲大 量如木瓜、胡蘿蔔…等食物，因含有豐富胡蘿蔔素導致色1259765 V. INSTRUCTION DESCRIPTION (1) [Technology of the Invention] The present invention utilizes a film image, which is controlled by a step-by-step control of a key area of white cold analysis to achieve a pre-study, a guide, a sink, or a turn. Deep stones, blood, including pumping, and internal vision to diagnose the technique] are common in the phenomenon of biliary red accumulation in tissues. Biliary obstruction of congenital blood and abdominal biliary and pancreatic surgery] Computer image processing technology, the clinical string, the moving photographic lens to dig the patient's eye; under the light source environment, with a unique color, two in the strict color, And the results of the analysis will be included in the S-cut and the split-bed diagnostic grading data and clinical fine-taste test data, and the accumulation of "Huangqi" symptoms in the laboratory will often occur in newborn children or hepatobiliary factors and their components. Appears = between, causing skin, white eyes and urine to change color. The causes of jaundice are liver disease, hepatitis, gallbladder, hemolysis, or other causes such as: side effects of drugs, Q-deficiency...etc. Clinically, the method of preliminary examination of jaundice is usually ultrasonic examination. If necessary, further liver puncture, computed tomography and magnetic resonance imaging are required. The oriental skin color is originally yellowish, so when the skin turns yellow, ', t 4 inch don't think about it, otherwise $ is neglected and not knowing. Usually the eye sclera changes the color of the skin in the patient's skin, so the key to the initial diagnosis is whether the color of the eyeball is yellow. At the same time, according to the clinician, the symptoms of mild plague are yellowing of the eyes and yellowing of the skin; when the symptoms of jaundice disappear, the color of the eyes is returned to normal. However, the doctor also pointed out that a large amount of food such as papaya, carrot, etc., is caused by rich carotene.
第5頁 1259765 五、發明說明⑵ 一— ^ ^，故頁疸病症診斷關鍵在於眼球鞏膜顏色是否變黃 判。不會與與飲食習慣相關之「胡、蘿葡素血症」混淆誤 關於眼球黃疸於診斷 醫方面 臟腑氣血盛衰、疾症i 邑澤，測知内在 :’縱觀目前市售非侵入性 另外於西请方 :測之「經皮測黃殖測:、雖：：，只有經皮膚 穩定。 之相性料及㈣”來得準確、 目前一般臨床醫師通當亩 眼球黃疫嚴重，，依直ί使？肉眼觀察來分級病患 的偏差，同時也不易伴在宙主 合易乂成人為主觀 麦此，本發明Γ;;::的臨；經驗值。 定，且存在有上述種種：缺‘知：：疸症狀之不易準確測 ，改良，遂得以首：=發：此潛心研究，並經多次 【發明内容】 本發明之目的在提供一 療檢測等技術’結合照明環 J學、影像處理及醫 軟體技術與臨床經驗資料康=/進驅動取像、影像 使用於醫療相關領域診；：考且=腦化檢測方法，以供 法。 啤，考之非知入式數位取樣比色方 具特徵在於.Λ 轉 之白色冷光板，以與提 _ _ τ 侍一可調光照 Λ 1徑制之识昍 •利用雙側可旋 之穩定白色光源Page 5 1259765 V. Description of the invention (2) One - ^ ^, so the key to the diagnosis of the disease is whether the color of the eye sclera becomes yellow. It will not be confused with the "hu and dyslipidemia" associated with eating habits. Mistakes about eyeball jaundice in the diagnosis of medical viscera, blood stasis, disease, i disease, internal knowledge: 'Overview of the current market non-invasive In addition, please contact us in the West: "Measurement of percutaneous test for yellowing: although::, only through the skin is stable. The phase material and (4)" is accurate, and the current general practitioners are serious in the eyeballs. Make? The deviation of the patients is classified by the naked eye, and it is also not easy to be accompanied by the main subject of the genus of the genus. This invention is 临;;:: 临; experience value. Ding, and there are all of the above: lack of 'know:: 疸 symptoms are not easy to accurately measure, improve, 遂 can be the first: = hair: this painstaking research, and many times [invention content] The purpose of the present invention is to provide a therapeutic test And other technologies 'in combination with lighting ring J, image processing and medical software technology and clinical experience data Kang = / drive image acquisition, imaging used in medical related field diagnosis;: test and = brain detection method, for the law. Beer, the non-intelligence digital sampling colorimetric formula is characterized by the 白色 turn white luminescent plate, with the _ _ τ servant adjustable light Λ 1 制 昍 利用 • use of bilateral rotatory stability White light source
第6頁 1259765 五、發明說明（3) ;B、步進驅動取 黃疫檢測，具自動 利用彩色影像處理 除眼球影像血管及 調分析之抽離方法 並省去使用昂貴之 庫：利用多次臨床 相關曲線，並將累 化黃疸分級之依據 【實施方法】 本發明之檢測 檢測作為實施例進 圍，請參照第一圖 A、 可控制之 可利用攝影取 第五圖所示，該攝 台（11)，該影像擷 該擷取平台之後方 該打燈箱（1 3 )可向 度以獲得最佳亮度 燈箱（1 3 )之光源使 ’亮度3100Lux)作 B、 步進驅動 在打燈箱（1 3 ) 方法採用 行說明， 所示，其 照明環境 像裝置： 影取像裝 取平台（1 處極設有 外樞轉開 ，避免產 用標準色 為眼球鞏 取像系統 之間的空 像系統：乃利用自動對焦取像裝置進行 杈位及對焦功能；c、影像軟體技術： 技術進行鞏膜關鍵區域之影像分割，去 眼角黏膜區域，再進行免校正之影像色 ，以簡化關鍵參數與黃疸之相依關係， 標準色板校正步驟。D、臨床經驗資料 人為診斷，蒐集關鍵參數與黃疸分級之 積經驗存入臨床資料庫，以便後續數據 〇 黃疸症狀之病患眼球鞏膜顏色 但不以此限定本發明之實施範 實施步驟如下： (1)為病人開始進行檢測，如 置（1)主要設有一影像擷取平 1 一）上設有一下巴托架（1 2 )，又 二呈夾角對置之打燈箱（1 3 )， 啟或者前後移動，調整照明角 生景> 像過暗或球體輝光。該打 溫白色冷光光源（色溫值5 3 〇 〇 κ 膜攝影光源。 間’裝設可供步進驅動之數位Page 6 1259065 V. Description of invention (3) ; B, stepper drive to take yellow plague detection, with automatic use of color image processing in addition to eyeball imaging blood vessels and analysis of the extraction method and save the use of expensive libraries: use multiple times Clinically relevant curve, and the basis for grading the jaundice [Method of Implementation] The detection test of the present invention is taken as an example, please refer to the first figure A, the controllable photographic can be taken as shown in the fifth figure, the camera (11), the image is after the drawing platform, the light box (13) can be used to obtain the best brightness light box (1 3) light source to make 'brightness 3100Lux' B, stepping drive in the light box ( 1 3) The method adopts the line description, as shown in the figure, the lighting environment is like the device: The image capturing and loading platform (the outer pole is turned away at one pole, avoiding the production of the standard color as the empty image between the eyeball image system) System: using the auto focus image capture device for clamping and focusing function; c, image software technology: technology to perform image segmentation in the key area of the sclera, go to the corner area of the mucous membrane, and then perform the image color without correction, to simplify The relationship between key parameters and Astragalus membranaceus, standard swatch correction procedure. D, clinical experience data for human diagnosis, the collection of key parameters and jaundice grading product experience stored in the clinical database, so that the subsequent data of jaundice symptoms of the eyeball sclera color but The implementation steps of the present invention are not limited to the following: (1) Starting the test for the patient, such as setting (1) mainly with an image capturing flat 1) having a chin bracket (1 2 ), and two The light box (1 3 ) with the opposite angle is set, open or move back and forth, adjust the illumination angle to produce a scene > like a dark or sphere glow. The warm white luminescent light source (color temperature value 5 3 〇 κ κ film photographic light source. The device is equipped with a digital drive for stepping drive)
、發明說明（4) _ 色攝影機(i4)，該彩色攝影機(⑷架設於移 1= 4如第六圖所示〕，該移動平台⑴)可帶動 (⑷之伸縮鏡頭；，並配合彩色攝影機 且有二产介門夕、ςί 方向移動，而彩色攝影機（14) 距！ 土間：；=，可針對病人(Α)之眼球進行焦 笙士同仏: （Α)之下巴靠置於下巴托架（12)上 軸：二L二則開始控制移動平台(15)呈χ軸及γ 位詈，^ 1 如乐八圖所示〕，以對準病人（A )之眼球, invention description (4) _ color camera (i4), the color camera ((4) is set to shift 1 = 4 as shown in the sixth figure), the mobile platform (1) can drive ((4) telescopic lens; and with color camera And there are two productions in the direction of the door, ςί, and the color camera (14) distance! Earth:; =, can be used for the eye of the patient (Α) to focus on the same: (Α) the chin rests on the chin The upper axis of the frame (12): the second L two starts to control the moving platform (15) and the y-axis and y-position 詈, ^ 1 as shown in the music diagram, to align the eye of the patient (A)
枝（4)之數位伸縮鏡頭，以作Z軸 之方向私動，而得以對準眼球之；t距。 軸 c、影像軟體技術 然後利用彩色攝影機（14)拍 將其所攝得之眼破旦彡#赵栌％ %二 之王早眼衫像，並 (B)的馬傻齡/ 象數據^虎轉換為紅⑴、綠（G)、藍 的〜像數位訊號，將取得眼 孤 球鞏膜部位之鲔&^ κ丰朕I顔色芰化，即眼 ^ M ^ , Hg 色數據輸入電腦進行分析處理，並於電 之…，員示。其實施步驟如 ：於， a•影像擷取： 卫間述如下· 利用數位化之攝影取像裝置進行取 置之結構容德句B日廿w 士 巧攝衫取像裝 像操取平台，然德在、史病心眼部#置於具下巴托架之影 號，並將：轉：ΑΡ ίί像取像裝置擷取眼球鞏膜影像訊 、隹^ \ 钇換為R、G 的數位訊號，暫存於雷向v 左之象時，攝影機之光圈約固定：5二 為3〇cm。 ，、?、度為3i〇〇Lux，而鏡頭至眼部之跑離約The digital telescopic lens of branch (4) is privately moved in the direction of the Z axis to be aligned with the eyeball; t distance. Axis c, image software technology and then use the color camera (14) to shoot the eye it is broken #赵栌%% 二王王早眼衫像, and (B) horse silly age / image data ^ tiger Converted to red (1), green (G), blue ~ like digital signal, will get the eye of the sclera of the eye and the ^ ^ κ Feng 朕 I color ,, that is, the eye ^ M ^, Hg color data input into the computer for analysis Processing, and in the electricity ..., the staff showed. The implementation steps are as follows: a, image capture: Weiwei said as follows: The structure of the digital image capture device is used to take the structure of the character of the sentence B Japanese 廿 w Shiqiao photo-taking image acquisition platform, Rand, in the history of the disease, the eye is placed in the shadow of the chin bracket, and will: turn: ΑΡ ίί image capture device to capture the eye sclera image, 隹 ^ \ 钇 changed to R, G digital signal When temporarily stored in the direction of the v to the left, the aperture of the camera is fixed: 5 is 3 〇 cm. ,,?, degree is 3i〇〇Lux, and the lens to the eye runs away from the appointment
1259765 五、發明說明（5) b .色調轉換： 在本發明中顏色的表示使用RGB模型〔紅（1〇、綠（G) 、藍（B)三分量〕，以及使用HSI模型〔色調（Hue)、飽和 度（Saturation)、亮度（Intensity) 三分量〕。擷取初 期所採用的色彩轉換方法是以國際照明協會（Comm i s s i on International de Γ E c 1 a i r a g e or International Commission on Illumination )C.I.E·標準色彩模型轉換 公式作為色彩轉換的輔助工具，利用關鍵色調之臨床經驗 累積，省卻標準色板校正之絕對值求取步驟。 c.影像分割： 瞳孔（黑）；我們 值依序分割皮 程區分成為影像 理之流程〔如第 ion Of 至H S I彩色模型 值分割、眼部與 最佳閥值分割及 影像血管及眼角 中心及計算出虹 影像後級分割處 級分割處理之後 Η、S)，資料庫 因為在影像處理分割的過程當中，分 依序為皮膚（黃）、鞏膜（白）、虹膜（棕）及 利用各區域色調間的相對差異，以不同閥 膚、鞏膜、虹膜及瞳孔。影像分割處理流 前、後級分割處理，其中影像前級分割處 三圖所示〕，依序為：選擇影像R(H (Reg Interest)，色彩對比伸張強化，RGB轉換 ’影像分割技術之皮膚與眼部影像最佳閥 蕈膜影像最佳閥值分割、鞏膜與虹膜影像 虹膜與瞳孔影像最佳閥值分割，去除眼球 黏膜區域，統計型態學處理之計算出瞳孔 膜半徑，影像二值化，映射回原影像。而 理〔如第四圖所示〕，依序為：於影像前 ，RGB色調轉換HIS，簡化關鍵參數值（r、1259765 V. INSTRUCTIONS (5) b. Tone conversion: In the present invention, the color representation uses the RGB model [red (1, green (G), blue (B) three components], and the HSI model [hue (Hue) ), saturation (Saturation), and brightness (Intensity) three components. The color conversion method used in the early stage of the acquisition is the International Lighting Association (Comm issi on International de Γ E c 1 airage or International Commission on Illumination) CIE standard The color model conversion formula is used as an auxiliary tool for color conversion, which accumulates the clinical experience of key tones, and eliminates the absolute value of the standard swatch correction. c. Image segmentation: pupil (black); Image processing flow (such as the first part of the ion of the color model segmentation, the eye and the best threshold segmentation and the image of the blood vessel and the corner of the eye and the calculation of the rainbow image after the segmentation process, S, S), database Because in the process of image processing and segmentation, the order is skin (yellow), sclera (white), iris (brown) and utilization of each area. The relative differences between the hues at different valve skin, sclera, iris and pupil. Image segmentation processing before and after the segmentation process, in which the image is divided into three parts, the sequence is: select image R (H (Reg Interest), color contrast stretch enhancement, RGB conversion 'image segmentation technology skin Optimal threshold segmentation of the optimal sacral membrane image with the eye image, optimal threshold segmentation of the iris and pupil images of the sclera and iris images, removal of the mucosal area of the eyeball, calculation of the pupillary membrane radius by statistical morphology processing, image binarity And map back to the original image. The rationale (as shown in the fourth figure), in order: before the image, RGB tone conversion HIS, simplify the key parameter values (r,
五、發明說明^6) " " ----—«— =關翏數曲線比對，黃疸數據分級，臨床醫師診斷修正， ^入^刪標準資料庫。最後再由分割出選定的虹膜中允π 二衫像中，分割出瞳孔區域，並以統計方法定位出瞳孔二 ，進—步計算出虹膜圓形半徑。最後以R = 215為閥值以 ^ '鞏膜影像血管區域，獲得最後所要分析最純淨之 d •黃疸參數比對： 建立黃疸顏色的等級標準，最初進行以彩色攝影機取 于眼球鞏膜影像顏色的r、G、B值後，再將各等級樣品之 腹j的平均值，透過統計平均計算而求出各級眼球愛 ^影像顏色平均的標準值。根據臨床實際檢測結果， :膜影像之R、H、S值與黃疸症狀之相關性最高，再各:;、、内 =二個標準值進行比較，可歸納到與標準值距離最接° 1 f級，最後樣本所屬的顏色就為三個等級中最大值。、，的 ^所求得之數據資料建立資料庫，予以輸入至亚將 作為檢測診斷時之判定標準。 内’以 D、臨床經驗資料庫 現行影像處理計算之數據資料，盥先前夫 ，診斷結果所建立之黃痕數據資料庫進行比； g達到黃疫症狀之認定標準’而提供作為醫療上 心:此將= = ί = :床診斷輔助檢 呈現，以協助醫事人員進行黃疫檢助 1259765 五、發明說明（7) 的精確度。 另，發 ，已製備相 作證明本發 故由以 點： 1. 本發 的創新工具 位影像處理 眼球特徵作 2. 本發 效，並可作 綜上所 且未見有相 能符合發明 明人為印證本案之可實施性及產業上之利用性 關實體樣品（實體照片如附件所示），經實體操 明確實可獲致上述之預期效果及實施目的。 上說明可知，本發明實施例確實具有下列之優 明可做為臨床上非侵入、免抽血診斷黃疸病情 ，不但減少抽血疼痛與傷口感染機會，經由數 的方法對鞏膜黃疸現象精確分級，而有助於對 定性及定量分析，提供臨床診斷依據。 明之檢測方法可提供早期發現，及早預防之功 為初期症狀之篩選，以防止繼續惡化。 述，本發明確實已具有可供產業上之利用性， 同發明之特徵公知、公用在先者，故本發明當 專利之申請要件、爰依法提出專利申請。 %V. Invention description ^6) "" -----«- = comparison of the number of curves, jaundice data classification, clinician diagnosis correction, ^ into the standard database. Finally, by dividing the selected iris in the selected iris, the pupil area is segmented, and the pupil is located by statistical method. The iris circular radius is calculated step by step. Finally, with R = 215 as the threshold to ^ 'scleral imaging vascular region, the last analysis to be the purest d • jaundice parameter comparison: establish the grading standard of jaundice color, initially take the color of the eye sclera image by color camera After the G and B values, the average value of the abdomen j of each grade sample is calculated by statistical average to obtain the standard value of the eyeball love image color average. According to the actual clinical test results, the correlation between the R, H, and S values of the membrane image and the symptoms of jaundice is the highest, and each of the two: standard values can be compared with the standard value. Level f, the color to which the last sample belongs is the maximum of the three levels. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , The data of the current image processing calculated by D, the clinical experience database, the ratio of the yellow mark data database established by the previous husband and the diagnosis result; g to achieve the standard of identification of the symptoms of the yellow plague' is provided as a medical center: This will = = ί = : bed diagnosis and auxiliary examination presentation, to assist medical staff to carry out the yellow plague inspection 1259765 V. The accuracy of the invention (7). In addition, the hair has been prepared to prove the origin of the accident by the point: 1. The innovative tool position of the hair image processing eyeball features 2. This effect, and can be used in summary and no phase can meet the invention Man-made verification of the feasibility of the case and the industrial use of the physical samples (physical photos as shown in the annex), through the actual gymnastics can achieve the above-mentioned expected effects and implementation purposes. As can be seen from the above description, the embodiments of the present invention have the following advantages as clinical non-invasive, blood-free diagnosis of jaundice, not only reducing the chance of blood-sucking pain and wound infection, but also accurately classifying the scleral jaundice phenomenon by a number of methods. It helps to provide a clinical diagnosis basis for qualitative and quantitative analysis. Ming's test method can provide early detection, early prevention is the screening of initial symptoms to prevent further deterioration. As described above, the present invention has indeed been available for industrial use, and the features of the invention are well known and publicly available. Therefore, the present invention is a patent application and a patent application according to law. %
第11頁 1259765 圖式簡單說明 第一圖為本發明之實施步驟流程圖。 第二圖為本發明影像軟體技術之步驟流程圖。 第三圖為本發明影像前級分割處理流程圖。 第四圖為本發明影像後級分割處理流程圖。 第五圖為本發明攝影取像裝置之外觀示意圖。 第六圖為本發明移動平台及彩色攝影機之組合示意圖。 第七圖為本發明使用時之示意圖。 第八圖為本發明移動平台之移動示意圖。 附件為本發明實體之實施照片。 【元件符號說明】 (1) 攝影取像裝置 (11) 影像擷取平台 (12) 下巴托架 (13) 打燈箱 (14) 彩色攝影機 (15) 移動平台 (A) 病人Page 11 1259765 BRIEF DESCRIPTION OF THE DRAWINGS The first figure is a flow chart of the implementation steps of the present invention. The second figure is a flow chart of the steps of the image software technology of the present invention. The third figure is a flow chart of the pre-image segmentation processing of the present invention. The fourth figure is a flowchart of the post-image segmentation processing of the present invention. The fifth figure is a schematic view of the appearance of the photographic image capturing apparatus of the present invention. The sixth figure is a schematic diagram of the combination of the mobile platform and the color camera of the present invention. The seventh figure is a schematic view of the invention when it is used. The eighth figure is a schematic diagram of the movement of the mobile platform of the present invention. The attached photo of the implementation of the entity of the invention. [Explanation of component symbols] (1) Photographic imaging device (11) Image capture platform (12) Chin tray (13) Light box (14) Color camera (15) Mobile platform (A) Patient