483747 A7 B7 5 05 8twf.doc/008 五、發明說明（f) 發明領域 (請先閱讀背面之注意事項再填寫本頁) 本發明是有關於一種資料與數位訊號的處理，且特別 是有關於一種運用呼吸系統狀況發出的聲音決定呼吸系統 狀況之方法與裝置。 習知技藝 呼吸系統的情況有著良好的文件記述，舉例來說，在 此處做爲參考的”兒童呼吸疾病之診斷與處理，，， G.M.Loughlin以及H.Eigen ’ 1994。大部分的呼吸聲都是 安靜地，所以當呼吸聲變爲聽得見的時候，通常就會引起 病人及其家屬的關心。例如喘鳴，氣喘，打鼾以及肺部阻 塞。另外在一些有著嚴重慢性支氣管炎、氣喘或者纖維性 囊腫的病人身上會聽到另一種雜音，這種雜音低沉而像漲 潮時隆隆作響，在離這種病人好幾呎之外就能聽見。 經濟部智慧財產局員工消費合作社印製 氣喘是一種連續性，幾乎像音樂般的聲音，通常在呼 氣時可以聽見。氣喘是空氣在大型及中型胸腔內的通氣道 中因爲流量限制導致擾動所產生。通過小型通氣道的氣體 流動率不足以造成氣喘。喘鳴是氣喘的一種型態，通常更 爲大聲而粗糙，一般主要在吸氣時聽到。它通常肇因於胸 腔外氣管內之阻礙物導致不斷增加的空氣擾動。打鼾是由 於咽喉軟體組織震動而造成的聲音。它通常是一種低沉、 只能在吸氣時聽見的喉音；除了一些極端的情形外，通常 只能在睡覺時聽見。肺部阻塞通常是病人用來槪括的一種 說辭。對於不同的人來說，它代表不同的意義，而通常它 被用來形容由有日漸增加的分泌物以及輕微縮小之主要通 4 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 483747 A7 B7 505 8twf.doc/〇〇8 五、發明說明（y) 氣道的未成年病患所造成的非特定聲塑。 有呼吸雜音的病人可以找內科醫生諮商。在這種諮商 的時候，第一步就是找到產生雜音的原由，然後給予一個 適當的建議來減輕症狀。然而，能用來評估呼吸雜音的選 項十分少。在多數的情況中，適切的評估需要前後方以及 側面胸腔X光照片。既然在很多情形之下，此雜音的來源 是良性的，那麼患者們就經常要在不必要的諮商之外接受 不必要的X光照相。 綜合說明 有鑒於此’本發明提出一種用來分析呼吸系統情況的 方法與裝置，它可以充分的避免因爲習知技藝的限制與缺 點所造成的多種問題。 本發明的其他特徵及優點一部份會揭露於接下來的敘 述中，其中一部份可以從敘述中明顯得知，或者可以從本 發明的實例中學得。本發明的目的及優點將可以藉由特別 在專利申請範圍中指出的功能及其組合獲得淸晰的認知。 爲了達成上述以及其他的優點，並且根據其中具體而 槪括論述的本發明之目的，這裡提供了一種根據由某情況 而造出的雜音來分析呼吸系統情況的方法。該方法包含了 計算由尙未確認的情況造出之雜音形成的數位訊號之參 數；以計算與參考模組一致的可能性來比較該參數與代表 呼吸系統中某一個已知狀況的參考模組；以及運用該參數 與參考模組的相似度來決定該參數是否與該參考模組吻 合’若吻合，則選擇此已知狀況來做爲該尙未確認的呼吸 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） (請先閱讀背面之注意事項再填寫本頁) m —ϋ ϋ I ϋ ϋ I 一# 口、I ^^1 ·ϋ I tmmmf 1^1 I · 經濟部智慧財產局員工消費合作社印製 A7 叫747 5〇58twf.doc/008 五、發明說明（> ) 系統狀況。 另外根據本發明，這裡提供了一種根據由某情況而造 出的雜音來分析呼吸系統情況的裝置。該裝置包含了由表 示尙未確認的情況所造出之雜音數位訊號中計算出參數的 裝置；以計算與參考模組一致的可能性來比較該參數與代 表呼吸系統中某一個已知狀況的參考模組的裝置；以及運 用該參數與參考模組的相似度來決定該參數是否與該參考 模組吻合’若吻合，則選擇此已知狀況來做爲該尙未確認 的呼吸系統狀況的裝置。 更進一步根據本發明來看，這裡提供了 一個可供電腦 讀取’包含了電腦軟體的媒體，當這個軟體在電腦上執行 的時候’能夠使得電腦成爲提供依據某未知情況而造出的 雜音來分析該呼吸系統之未知情況的工具。這個工具包含 了由表示尙未確認的情況所造出之雜音的格式化數位訊號 中計算出參數的裝置；以計算與參考模組的相似度來比較 該參數與代表呼吸系統中某一個已知狀況的參考模組的裝 置；以及運用該參數與參考模組的相似度來決定該參數是 否與該參考模組吻合，若吻合，則選擇此相對應的已知狀 況來做爲該尙未確認的呼吸系統狀況之裝置。 需要注意的是，以上所做的敘述以及接下來的詳細敘 述兩者都只是示範而且有助於說明，並不是如申請專利範 圍一樣成爲對本發明的一種限制。 式說明 隨同附上結合並組成這份申請書其中一部份的圖式， 6 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） -----，------裝--------訂---------^9— (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製 483747 5058twf.doc/008 A7 B7 五 濟 部 智 慧 財 員 工 消 費 合 作 製 發明說明（4) 乃是展示出本發明的一個實例，而且跟下列敘述一起來解 釋本發明的目的、優點以及原理，在這些圖形中·· 第1圖繪示的是人類呼吸系統的模型； 第2圖繪示的是以普通方塊圖表示，根據本發明所建 立的一個決定人類呼吸系統狀況之示範系統； 第3圖繪示的是一個示範流程圖，該圖描述的是有關 根據本發明，從一狀況所得的雜音來決定呼吸系統狀況之 步驟； 第4圖繪示的是一個示範流程圖，該圖描述的是有關 根據本發明來計算所得參數之步驟； 第5圖繪示的是一個示範流程圖，該圖描述的是有關 根據本發明，將至少一個參數與至少一個參考模組加以比 較的步驟；以及 第6圖繪示的是一個示範流程圖，該圖描述的是有關 根據本發明來選擇一種狀況的步驟。 主要_元件編號 100 :呼吸系統 105 :肺 110 :聲帶 120 :喉管 125 :咽腔 127 :軟顎 129 :舌頭 130 :鼻腔 7 (請先閱讀背面之注意事項再填寫本頁)483747 A7 B7 5 05 8twf.doc / 008 V. Description of the invention (f) Field of invention (please read the notes on the back before filling out this page) The present invention relates to the processing of data and digital signals, and in particular to A method and device for determining the condition of the respiratory system by using sounds emitted by the condition of the respiratory system. The state of the art of the respiratory system is well documented. For example, "Diagnosis and Management of Respiratory Diseases in Children," GMLoughlin and H. Eigen '1994, referenced here. Most of the breathing sounds are It is quiet, so when the sound of breathing becomes audible, it usually causes concern for patients and their families. For example, wheezing, asthma, snoring, and lung obstruction. In addition, some people have severe chronic bronchitis, asthma, or fiber. Patients with sexual cysts will hear another kind of murmur, which is low and rumbling like high tide, and can be heard from a few feet away from such patients. The Ministry of Economic Affairs, Intellectual Property Bureau, Employee Consumer Cooperative, printed asthma is a continuous Sexual, almost music-like sound, usually audible during exhalation. Asthma is caused by disturbance of air in the airways in large and medium chests due to flow restrictions. The flow rate of gas through small airways is not sufficient to cause asthma . Wheezing is a form of asthma that is usually louder and rougher and is generally heard mainly during inhalation. It Often due to increasing air disturbances caused by obstructions in the trachea outside the chest. Snoring is a sound caused by the vibration of the throat soft tissue. It is usually a low-pitched larynx that can only be heard during inspiration; except for some extremes It is usually only heard during sleep. Pulmonary obstruction is often used by patients as a form of expression. For different people, it means different meanings, and it is usually used to describe the increasing The secretions and slightly reduced main contents 4 The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 483747 A7 B7 505 8twf.doc / 〇〇8 V. Description of the invention (y) Non-specific vocal plaques caused by adult patients. Patients with respiratory murmur can consult a physician. In this consultation, the first step is to find the cause of the murmur and then give an appropriate recommendation to reduce the symptoms However, there are very few options available for assessing respiratory murmur. In most cases, proper assessment requires anterior and posterior and lateral chest x-rays. In many cases, the source of this murmur is benign, so patients often need to receive unnecessary X-ray photography in addition to unnecessary consultations. Comprehensive explanation In view of this, the present invention proposes a method for analyzing the respiratory system The method and device of the situation can fully avoid various problems caused by the limitations and disadvantages of the know-how. Part of the other features and advantages of the present invention will be disclosed in the following description. It is obvious from the description or can be learned from the examples of the present invention. The objects and advantages of the present invention can be obtained through a clear understanding of the functions and combinations specifically pointed out in the scope of the patent application. In order to achieve the above and other Advantages, and in accordance with the purpose of the present invention which is specifically and comprehensively discussed herein, a method for analyzing the condition of the respiratory system based on the noise generated by a certain situation is provided here. The method includes calculating a parameter of a digital signal formed by noise generated by the unconfirmed situation; comparing the parameter with a reference module representing a known condition in the respiratory system by calculating the possibility of consistency with the reference module; And use the similarity between the parameter and the reference module to determine whether the parameter is consistent with the reference module. If it matches, then select this known condition as the unconfirmed breath. The paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) (Please read the precautions on the back before filling in this page) m —ϋ ϋ I ϋ 一 I 一 # 口 、 I ^^ 1 · m I tmmmf 1 ^ 1 I · Ministry of Economy Wisdom The A7 printed by the Property Cooperative Consumer Cooperative is called 747 5〇58twf.doc / 008 5. Description of the invention (>) System status. In addition, according to the present invention, there is provided a device for analyzing a condition of a respiratory system based on a noise generated by a certain condition. The device includes a device for calculating a parameter from a digital signal of noise generated by an unconfirmed condition; comparing the parameter with a reference representing a known condition in the respiratory system by calculating the possibility of consistency with the reference module The device of the module; and using the similarity of the parameter with the reference module to determine whether the parameter agrees with the reference module. If it matches, then select this known condition as the device of the unconfirmed respiratory system condition. According to the present invention, here is provided a computer-readable 'media containing computer software, when this software is run on a computer' can make the computer to provide noise generated according to an unknown situation A tool to analyze unknown conditions of the respiratory system. This tool contains a device that calculates a parameter from a formatted digital signal representing noise generated by an unconfirmed condition; compares the parameter to a known condition in the respiratory system by calculating the similarity with a reference module Device of the reference module; and using the similarity between the parameter and the reference module to determine whether the parameter agrees with the reference module, and if it matches, select the corresponding known condition as the unconfirmed breath Device for system status. It should be noted that both the above description and the following detailed description are merely exemplary and helpful for explanation, and are not a limitation on the present invention as the scope of patent application. A description of the formula is attached together with the drawings that form part of this application. 6 paper sizes are applicable to the Chinese National Standard (CNS) A4 (210 X 297 mm) -----, ---- --Installation -------- Order --------- ^ 9— (Please read the notes on the back before filling this page) Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 483747 5058twf. doc / 008 A7 B7 The description of the invention of the consumption cooperation system of the employees of the Five-Family Ministry of Wisdom (4) is an example of the present invention, and together with the following description to explain the purpose, advantages and principles of the present invention, in these figures · Figure 1 shows a model of the human respiratory system; Figure 2 shows a typical block diagram, an exemplary system established in accordance with the present invention to determine the condition of the human respiratory system; Figure 3 shows a An exemplary flowchart depicting the steps involved in determining the condition of the respiratory system from the noise obtained from a condition according to the present invention; Figure 4 illustrates an exemplary flowchart depicting the steps related to the present invention. Steps to calculate the obtained parameters; Figure 5 illustrates an exemplary flowchart illustrating the steps involved in comparing at least one parameter with at least one reference module according to the present invention; and Figure 6 illustrates an exemplary flowchart illustrating The figure describes the steps involved in selecting a situation in accordance with the present invention. Main_component number 100: respiratory system 105: lung 110: vocal cord 120: larynx tube 125: pharyngeal cavity 127: soft palate 129: tongue 130: nasal cavity 7 (Please read the precautions on the back before filling this page)
本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 483747 A7 B7 505 8twf.doc/008 五、發明說明（f) 140 : 口腔 150 :聲音 205 :輸入模組 210 :類比/數位轉換模組 220 :資料處理器 230 :儲存模組 240 :輸出模組 242 :顯示裝置 244 :印表機 眚施例 現在我們將詳細說明根據本發明而來的較佳實施例， 這些例子繪示在附圖中。我們將會盡可能對圖中一致或相 似的部分使用同樣的參考號碼。 第1圖所畫的是一個人類呼吸系統模型100。從第1 圖中可以看出，這個人類呼吸系統100包含了肺105，聲 帶110，喉管120，咽腔125，軟顎127，舌頭129，鼻腔 130以及口腔140。呼吸系統100製造出來獨特的聲音！50 是在呼吸系統狀況裡的典型。、表1所列舉的是根據本發 明，能加以診斷的呼吸系統可能之狀況。 雖然一個受過訓練的人可能可以區分一些根據某狀況 產生的聲音，但是大部分的人非常難以僅根據聲音就能診 斷狀況。經由利用接收到之代表呼吸系統狀況的聲音，本 發明提供了一個能指出呼吸系統可能狀況的的診斷輸出， 並且在許多實例中，可以消除不必要之X光照射或是醫藥 _本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） (請先閱讀背面之注意事項再填寫本頁) rl裝--------訂--------- 經濟部智慧財產局員工消費合作社印製 483747 5 05 8twf.doc/008 A7 B7 五、發明說明（乙 諮商的需求 1 過敏反應 2過敏性支氣管與肺麴菌病20 經濟部智慧財產局員工消費合作社印製 3 血管淋巴腺水腫 4 呼吸-吞嚥功能不良 5 氣喘⑴ 6 氣喘(II) 7 纖維性囊腫 8 會厭炎 9 會厭囊腫 10 異物阻塞 11 噴出氣流逆流 12 喉部腫囊，血管瘤 13 喉頭氣管與支氣管炎 14 喉頭氣管食道腺破裂 15 喉頭軟化 16 喉頭乳狀瘤 17 縱隔節點或區塊 18 咽部潰瘍 表1.呼吸系統之狀況 19 血鐵質侵害肺部 存在異物(I) 存在異物(II) 存在異物(III) 突發假膜性喉炎 次聲門狹窄 呼吸道狹窄 氣管軟化 心理性休克 血管鳴響(vascular ring) 病毒性細支氣管炎 聲帶功能不良 初發打鼾 障礙性睡眠呼吸暫停 鼻音-鼻後孔狹窄 鼻音-鼻咽炎 支氣管-嘯喘 支氣管炎（傳染性過敏 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 阻塞） 第2圖所畫的是一個根據本發明所建構的示範系統 2〇〇，該系統用來決定一個人類呼吸系統的狀況。如在第2 •裝--- (請先閱讀背面之注意事項再填寫本頁) 0 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 483747 A7 B7 505 8twf.doc/008 五、發明說明（q) (請先閱讀背面之注意事項再填寫本頁) 圖中所看到的，這個系統200包含了一個輸入模組205， 一個類比/數位(A/D)轉換模組210，一個資料處理器220， 一個儲存模組230以及一個輸出模組240。 經濟部智慧財產局員工消費合作社印製 表示在第2圖系統200中的輸入模組205，接收從呼 吸系統1〇〇之狀況產生的聲音150來分析該狀況。這個輸 入模組205可以由各種能夠輕易取代以及完成提供輸入到 此類比到數位轉換模組210的裝置構成。舉例來說，這些 裝置其中一些包括了，網路介面模組，一般功能的數位介 面模組以及麥克風。雖然像第2圖所示的A/D轉換模組會 用在某些場合中，但是若聲音150早在提供到系統200之 前就已經轉換成數位型式，那就不需要A/D轉換模組了。 對於熟悉此技藝的人而言，顯而易見的這個資料處理器可 能包含了例如一些額外的硬體，軟體以及/或類似一組中央 處理單元，輔助處理器，記憶體，暫存器以及其他處理的 裝置和系統等合適的韌體。儲存模組230可以由一群多變 的元件或次系統所組成，舉例來說，包括硬碟驅動裝置， 光碟驅動裝置，一般功能的儲存裝置，可拆式儲存裝置以 及其他可以用來完成儲存功能的裝置。輸出模組230可以 由一群多變的系統、次系統以及元件所組成，舉例來說， 包括顯示裝置242以及印表機244。顯示裝置242可以由 一群多變的裝置所組成，舉例來說，包括陰極射線管或者 液晶顯示器。除此之外，此輸出裝置240可以包括其他能 提供輸出功能的系統、副系統以及裝置，例如網路介面， 儲存裝置，一般功能的數位介面，平行介面以及串聯介面。 10 ^紙張尺度適用中國國家標準（CNS)A4規格（21〇 X 297公釐） 483747 A7 B7 5 05 8twf.doc/008 五、發明說明（3 ) 此外，此系統200以及/或任一上述的模組，裝置，元件或 功能可以被含括於硬體、軟體或者韌體的任意合適組合 中，而且可以當作一個標準獨立系統或是與其他裝置合而 爲一。 輸入模組205接收由呼吸系統100發出的聲音150。 在第2圖的示範實施例中，該輸入模組205包括了一個麥 克風來產生一組代表聲音150的類比聲音訊號。此A/D轉 換模組接收從輸入模組205來了類比聲音訊號，並且將該 類比聲音訊號轉化成數位訊號。資料處理器220處理該數 位訊號以產生一組格式化的數位訊號，計算至少一個代表 聲音150的參數，比較該至少一個的參數與至少一個對應 於列表在表格1中，儲存在該儲存模組230之狀況的參考 模組，並且決定表示聲音150的可能狀況。該輸出模組240 提供一個狀況輸出，以及選擇性的輸出-應付此狀況之醫藥 指示。 第3圖表示的是一個示範的流程圖，上面畫有一組關 於利用由某狀況產生的聲音150來決定呼吸系統1〇〇之狀 況的程序300相關步驟。如第3圖所示，該程序包含了第2 圖中系統200實施的步驟。程序300開始於系統100產生 被輸入模組接收的聲音150。該輸入裝置205將聲音150 轉換成一組類比聲音訊號。 爲了將該聲音訊號轉換成一組格式化的數位訊號，該 A/D轉換模組接收該類比聲音訊號(步驟310)，對該接收到 的類比聲音訊號進行取樣(步驟320)，並且將該類比聲音訊 ------^------裝--------訂---------. (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 483747 5 05 8twf. doc/008 A7 B7 經濟部智慧財產局員工消費合作社印製 五、發明說明（巧） 號轉換爲數位訊號（步驟330)。在一個示範的實施例中，該 A/D轉換模組210以大約每秒16000個取樣的取樣率對類 比聲音訊號進行了期間大約20秒的取樣。在該示範實施例 中，這20秒的類比聲音訊號經由轉換，產生了一群足夠做 出正確分析的數位訊號。然而，品質差的類比聲音訊號可 能需要更大群的數位訊號（也就是說，更快的取樣率或是更 長的取樣期間）。雖然第3圖表示了接收一組類比聲音訊號 (步驟310)，該系統200可以經由從步驟335開始而輕易的 接受代表該狀況之類比聲音訊號的數位訊號。 爲了產生該格式化的數位訊號，該資料處理器220處 理了該數位訊號（步驟335)。一種適當的格式化數位訊號結 構可以提供最棒的處理表現以及精確的診斷。雖然其他種 類的結構或格式可能被採用，但是在該示範實施例中，該 資料處理器220將該數位訊號切割成一連串的區塊（步驟 3；35)，每一個區塊期間爲25毫秒，在連串區塊間插有1〇 毫秒的重疊資料以提供流暢的資料型態。 該資料處理器220使用該格式化的數位訊號計算至少 —個參數(步驟34〇)，第4圖是一個示範的流程圖，所畫的 是根據本發明來計算參數(步驟34〇)的程序400之步驟。如 第4圖所示，該資料處理器220對格式化的數位訊號區塊 進行局速傅立葉轉換(步驟410)。得到的頻域値經過對數轉 換(步驟420)以及一次反高速傳立葉轉換(FFT-1)(步驟43〇) 來提供做爲該至少一個參數的一組係數。這些係數又稱爲 頰頻率係數（mel-frequency)。雖然該示範實施例使用了頰 12 (請先閱讀背面之注意事項再填寫本頁) 裝---- 訂---------線| 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 483747 5058twf.doc/008 A7 B7 經濟部智慧財產局員工消費合作社印製 五、發明說明（（b) 頻率係數，我們仍然可以用其他的技術來計算這些係數， 舉例來說，包括次頻帶輸出（sub-band outputs)，線性頻譜 對（line spectrum pairs)以及知覺性加重線性預測係數 (perceptually weighted linear prediction)。在語音辨識領域 中，我們將”語音辨識之基礎”，L.Rabiner以及B.Juang， B93加入在此做爲參考，其中描述了頰頻率係數，次頻帶 輸出，線性頻譜對以及知覺性加重線性預測係數的計算。 重新參考第3圖，一些該組係數被用來當作該至少一 個參數，而該參數被拿來與至少一個儲存於儲存模組230， 表示系統100可能經歷的該些可能呼吸狀況的參考模組相 比較(步驟350)。 該參數計算(步驟340)產生至少一個參數(也就是，至 少一個値）。該資料處理器220需要至少一個參數來與至少 一個參考模組相比較(步驟350)。爲比較(步驟350)所需的 最佳參數數目是不固定的。若是使用了太少的參數，則由 輸出模組130提供的最後診斷就會比較不精確。相反的， 若是使用了過多的參數，則在資料處理器220中所需的處 理以及儲存將會變得難以負擔。在一個示範實施例中，該 創作人對39個參數(也就是39個値)與該至少一個的參考 模組做了比較。 弟5圖繪不的是一'個不範流程圖，該圖描述的是包含 有關根據本發明，將至少一個參數與至少一個參考模組加 以比較之副程序的程序500。根據第5圖，該資料處理器 220用計算該至少一個參數與該至少一個參考模組（步驟 13 請 閱 讀 背，This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 483747 A7 B7 505 8twf.doc / 008 V. Description of the invention (f) 140: Oral 150: Sound 205: Input module 210: Analog / Digital conversion module 220: data processor 230: storage module 240: output module 242: display device 244: printer. Examples Now we will explain in detail the preferred embodiments according to the present invention. These examples are drawn Shown in the drawing. Wherever possible, we will use the same reference numbers for identical or similar parts of the figures. Figure 1 depicts a human respiratory system model 100. As can be seen from Figure 1, the human respiratory system 100 includes a lung 105, a vocal cord 110, a throat 120, a pharyngeal cavity 125, a soft palate 127, a tongue 129, a nasal cavity 130, and an oral cavity 140. The breathing system 100 makes unique sounds! 50 is typical in respiratory conditions. Table 1 lists possible conditions of the respiratory system that can be diagnosed according to the present invention. Although a trained person may be able to distinguish some sounds based on a certain condition, it is very difficult for most people to diagnose the condition based on only sound. The present invention provides a diagnostic output that indicates possible conditions of the respiratory system by using sounds that are representative of the condition of the respiratory system, and in many instances, eliminates unnecessary X-ray exposure or medicine. This paper applies China National Standard (CNS) A4 Specification (210 X 297 mm) (Please read the precautions on the back before filling out this page) rl Pack -------- Order --------- Ministry of Economic Affairs Printed by the Intellectual Property Bureau employee consumer cooperative 483747 5 05 8twf.doc / 008 A7 B7 V. Description of the invention (B Consultant's requirements 1 Allergic reactions 2 Allergic bronchial and pulmonary bacillary disease 20 Control 3 Angiolymphatic gland edema 4 Respiratory-swallowing dysfunction 5 Asthma ⑴ 6 Asthma (II) 7 Fibrous cysts 8 Epiglotitis 9 Epiginal cysts 10 Foreign body obstruction 11 Spit airflow countercurrent 12 Laryngeal cysts, hemangioma 13 Laryngeal trachea and bronchus Inflammation of the laryngeal tracheoesophageal glands 15 Softening of the throat 16 Laryngeal papillomas 17 Mediastinal nodes or blocks 18 Ulcers of the pharynx (II) Presence of foreign body (III) Sudden pseudomembranous laryngitis Subglottic stenosis Respiratory stenosis Airway softening Psychological shock Vascular ringing Viral bronchiolitis Vocal dysfunction Initial snoring disorder Sleep apnea nasal sound- Nasal narrow nasal sound-nasopharyngitis bronchus-wheezing bronchitis (infectious allergy 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 obstruction) Figure 2 shows a structure constructed according to the present invention Demonstration system 2000, which is used to determine the condition of a human respiratory system. For example, in the second installation (Please read the precautions on the back before filling out this page) 0 This paper size applies to Chinese national standards (CNS ) A4 specification (210 X 297 mm) 483747 A7 B7 505 8twf.doc / 008 V. Description of the invention (q) (Please read the notes on the back before filling this page) As you can see in the figure, this system 200 contains An input module 205, an analog / digital (A / D) conversion module 210, a data processor 220, a storage module 230, and an output module 240. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs The input module 205 shown in the system 200 in FIG. 2 receives a sound 150 generated from the condition of the respiratory system 100 to analyze the condition. This input module 205 can be easily replaced by various types and can provide input to such a comparison. The device configuration to the digital conversion module 210. For example, some of these devices include a network interface module, a general-purpose digital interface module, and a microphone. Although the A / D conversion module as shown in Figure 2 will be used in some occasions, if the sound 150 has been converted into a digital type before it is provided to the system 200, then the A / D conversion module is not required Already. For those skilled in the art, it is obvious that the data processor may include, for example, some additional hardware, software, and / or a similar set of central processing units, auxiliary processors, memory, registers, and other processing Suitable firmware such as devices and systems. The storage module 230 may be composed of a group of changeable components or sub-systems, for example, including hard disk drive devices, optical disk drive devices, general-purpose storage devices, removable storage devices, and other storage functions that can be used to complete storage functions. installation. The output module 230 may be composed of a group of changeable systems, subsystems, and components. For example, the output module 230 includes a display device 242 and a printer 244. The display device 242 may be composed of a group of changeable devices, including, for example, a cathode ray tube or a liquid crystal display. In addition, the output device 240 may include other systems, sub-systems, and devices capable of providing output functions, such as a network interface, a storage device, a general-purpose digital interface, a parallel interface, and a tandem interface. 10 ^ Paper size applies Chinese National Standard (CNS) A4 specification (21 × 297 mm) 483747 A7 B7 5 05 8twf.doc / 008 V. Description of invention (3) In addition, this system 200 and / or any of the above A module, device, component, or function can be included in any suitable combination of hardware, software, or firmware, and can be considered as a standard stand-alone system or integrated with other devices. The input module 205 receives a sound 150 emitted by the respiratory system 100. In the exemplary embodiment of FIG. 2, the input module 205 includes a microphone to generate a set of analog sound signals representing the sound 150. The A / D conversion module receives an analog sound signal from the input module 205 and converts the analog sound signal into a digital signal. The data processor 220 processes the digital signals to generate a set of formatted digital signals, calculates at least one parameter representing the sound 150, compares the at least one parameter with at least one corresponding to the list in Table 1, and stores in the storage module The reference module of the status of 230, and decides to indicate the possible status of the sound 150. The output module 240 provides a status output and a selective output-a medical indication to cope with the status. Figure 3 shows an exemplary flow chart with a set of steps related to a procedure 300 for determining the condition of the respiratory system 100 using the sound 150 generated by a condition. As shown in FIG. 3, the program includes steps implemented by the system 200 in FIG. The process 300 begins when the system 100 generates a sound 150 received by the input module. The input device 205 converts the sound 150 into a set of analog sound signals. In order to convert the sound signal into a set of formatted digital signals, the A / D conversion module receives the analog sound signal (step 310), samples the received analog sound signal (step 320), and the analog signal Voice message ------ ^ ------ install -------- order ---------. (Please read the notes on the back before filling this page) Ministry of Economy Printed by the Intellectual Property Bureau Staff Consumer Cooperatives This paper is scaled to the Chinese National Standard (CNS) A4 (210 X 297 mm) 483747 5 05 8twf. Doc / 008 A7 B7 Printed by the Intellectual Property Bureau Employee Consumer Cooperatives of the Ministry of Economic Affairs The description signal is converted into a digital signal (step 330). In an exemplary embodiment, the A / D conversion module 210 samples the analog audio signal for a period of about 20 seconds at a sampling rate of about 16,000 samples per second. In this exemplary embodiment, the 20-second analog sound signal is converted to produce a group of digital signals sufficient for proper analysis. However, poor quality analog audio signals may require a larger group of digital signals (that is, faster sampling rates or longer sampling periods). Although FIG. 3 shows receiving a set of analog audio signals (step 310), the system 200 can easily accept digital signals representing the analog audio signals representing the situation, starting from step 335. To generate the formatted digital signal, the data processor 220 processes the digital signal (step 335). A properly formatted digital signal structure provides the best processing performance and accurate diagnostics. Although other types of structures or formats may be used, in the exemplary embodiment, the data processor 220 cuts the digital signal into a series of blocks (step 3; 35), each block period is 25 milliseconds, 10 milliseconds of overlapping data are interpolated between a series of blocks to provide a smooth data pattern. The data processor 220 uses the formatted digital signal to calculate at least one parameter (step 34). FIG. 4 is an exemplary flowchart illustrating the procedure for calculating parameters (step 34) according to the present invention. Step 400. As shown in FIG. 4, the data processor 220 performs local speed Fourier transform on the formatted digital signal block (step 410). The obtained frequency domain 値 is subjected to a logarithmic transformation (step 420) and an inverse high-speed Fourier transform (FFT-1) (step 43) to provide a set of coefficients as the at least one parameter. These coefficients are also called mel-frequency coefficients. Although this exemplary embodiment uses cheek 12 (please read the precautions on the back before filling this page) Binding ---- Order --------- line | This paper size applies to China National Standard (CNS) A4 Specifications (210 X 297 mm) 483747 5058twf.doc / 008 A7 B7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 5. Description of the invention ((b) Frequency coefficients, we can still use other techniques to calculate these coefficients, for example For example, it includes sub-band outputs, line spectrum pairs, and perceptually weighted linear prediction. In the field of speech recognition, we will "basics of speech recognition" , L.Rabiner and B.Juang, B93 are added here as a reference, which describes the calculation of buccal frequency coefficients, sub-band output, linear spectral pairs, and the linear prediction coefficients of perceptual weighting. Refer back to Figure 3, some of this group The coefficient is used as the at least one parameter, and the parameter is used with at least one stored in the storage module 230, indicating the possible breathing conditions that the system 100 may experience. The reference module is compared (step 350). The parameter calculation (step 340) generates at least one parameter (ie, at least one 値). The data processor 220 needs at least one parameter to compare with at least one reference module (step 350). The optimal number of parameters required for comparison (step 350) is not fixed. If too few parameters are used, the final diagnosis provided by output module 130 will be less accurate. Conversely, if used With too many parameters, the processing and storage required in the data processor 220 will become unaffordable. In an exemplary embodiment, the creator has 39 parameters (that is, 39 与) and the at least one The reference module is compared. The figure 5 does not depict a 'non-standard flow chart', which depicts a subroutine for comparing at least one parameter with at least one reference module according to the present invention. Procedure 500. According to FIG. 5, the data processor 220 calculates the at least one parameter and the at least one reference module (step 13 please read the back,
意 事 項 再 f裝 頁I 訂 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公爱） 483747 A7 B7 5 05 8twf. doc/008 五、發明說明（（（） 520)的機率或相似度的値(步驟510)來作比較（第3圖之步 驟350)，該機率或相似度的値就是該參考模組(步驟520) 和從該格式化的數位訊號計算出來的該至少一個參數的相 似程度。該計算出來的相似度的値提供了一該至少一個參 數與該相對應的參考模組間的相似程度之相關量測。在本 發明的一個示範實施例中，該資料處理器220使用了一個 麥特比解碼器（Viterbi Decoder)來計算相似的機率（步驟 510)。在一個示範實施例中，該資料處理器220重複(步驟 530)對任一儲存於儲存模組230的36種參考模組計算一機 率或者相似度的値(步驟510)，其中該36種參考模組各對 應於列在表1中的3 6種呼吸系統之狀況。雖然這個實施例 使用了麥特比解碼器，我們仍然可以使用其他的技術來比 較該至少一個參數與該至少一個參考模組，舉例來說，包 括動態時間扭曲法（dynamic time warping method)，向量量 子化（vector quantization)，或是人工神經網路（artificial neural networks)。在語音辨識領域中，”語音辨識之基礎”， L.Rabiner 以及 B.Juang，1993，敘述了麥特比解碼（Viterbi Decode)，動態時間扭曲，向量量子化以及人工神經網路。 重新參考第3圖，在本發明的一個實施例中，該資料 處理器220使用了一做爲要分析之呼吸系統的已知狀況訓 練樣本，計算一表示該呼吸系統1〇〇的一以知狀況的參考 模組。該訓練樣本大體上由該呼吸系統之已知狀況產生的 聲音所組成。爲了組成該參考模組，該A/D轉換模組將該 訓練樣本轉換成數位訊號，而該資料處理器22〇將該數位 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公爱） (請先閱讀背面之注意事項再填寫本頁) 1 am— ·1 ϋ i_i I 一 ον ϋ ·ϋ ϋ ϋ I ϋ I · 經濟部智慧財產局員工消費合作社印製 483747 A7 B7 5058twf.doc/008 五、發明說明（（l) 訊號格式化以提供一個格式化的數位訊號，並且計算出至 少一個參數，作成要被診斷的已知情況之相對應參考模組 (步驟330，335，340)。結果，該參考模組包括了至少一個 參數，該至少一個參數對應於保留在儲存模組230中要被 診斷的已知狀況的訓練樣本。對於任一個要被診斷的其他 已知狀況，該資料處理器220使用一用在該其他已知狀況 的其他訓練樣本，以計算表示該其他呼吸系統已知狀況的 其他參考模組。 在本發明的一個實施例中，該儲存裝置230包含了根 據每一被分辨以及診斷的狀況，使用隱藏式馬可模組 (Hidden Markov model)計算出的至少一個參考模組。該資 料處理器220利用每一個呼吸系統狀況的訓練樣本以計算 隱藏式馬可模組。舉例來說，該資料處理器220利用了根 據該訓練樣本得到的格式化數位訊號，計算出該狀況對應 於隱藏式馬可模組的參考模組。”語音辨識之基礎”， L.Rabiner以及B.Juang，1993，敘述了運用於語音辨識中 的隱藏式馬可模組。 參考第3圖，爲了決定該格式化的數位訊號是否與該 參考模組相配，該資料處理器220決定(步驟360)對應於該 參考模組之一的該至少一個參數，或者一組參數，是否在 一個可接受的相似値範圍內。更詳細的說，該資料處理器 220從儲存於儲存模組230,以至少一個參考模組表示的呼 吸系統100情況之中做出一個決定以及選擇（步驟360)。做 爲決定以及選擇一可能情況的一部份，在提供一輸出前， 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公爱） (請先閱讀背面之注意事項再填寫本頁) 【裝--------訂---------線. 經濟部智慧財產局員工消費合作社印製 483747 A7 B7 5 05 8twf.doc/0〇8 五、發明說明（A) 該資料處理益220只選擇，或說比較喜歡，最相似的情形， 而過濾掉不能接受或不相似的情況。 第6圖表示的是一個程序600之示範流程圖，該程序 包含了有關根據本發明來選擇一種狀況的步驟。參考第6 圖，在根據本發明的一個示範實施例中，該資料處理器220 使用表示在程序600中的步驟來決定並選擇狀況（步驟 360)。該資料處理器22〇使用了從該比較步驟（步驟35〇) 接收:到的可能性來評估各狀況相對應的可能性（步驟 61〇)。該資料處理器22〇選擇了五個最大可能性的値（步驟 62〇) °假如該最大五個値有一個可接受的可能性的値(步驟 63〇) ’則該資料處理器22〇就提供該狀況來當輸出（步驟 650 ’ 660) °假如該五個狀況裡的任一個都不可能代表該問 題的真實狀況，該資料處理器220就只選擇該些可接受的 并犬況（步驟640)，並且提供一個適當的輸出（步驟650，660) 到輸出模組240。該輸出可以包含有針對該（些）選擇情形的 詳細醫藥資訊（步驟66〇)。如同在流程圖中提及的，輸出可 以由該輸出模組以一個方便的模式顯示給一使用者（步驟 650) 〇 再参考第3圖，該輸出模組240提供一狀況的指示給 一使用者(步驟370)。在一示範實施例中，該輸出模組包括 了，舉例來說，一個顯示器^42以及一個印表機244，以 提供〜使用者對應於該狀況的詳細醫藥資訊(步驟66〇)。 因此’本發明提供了新的裝置以及方法來診斷呼吸系 統的狀況。 (請先閱讀背面之注意事項再填寫本頁) -裝--------訂---------線j 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準（CNS)A‘！規格mo X 297公爱） 483747 5058twf.doc/008 五、發明說明（（u) 本發明之較佳實施例已揭露如上，目的在於表示及欽 述其彳示的。适並不是要徹底g羊盡欽述或是限制本發明於該 揭露之精確的形式，而且經由以上的教學啓示或是本發明 的實行，可能可以做出修改以及多樣化。舉例來說，本發 明並不限定於揭露在第1圖中的裝置，而是一般可以各種 方法來實施的，本發明之範圍定義在申請專利範圍以及其 同等意義之中。 任何熟習此技藝者，在不脫離本發明之精神和範圍 內，當可作各種之更動與潤飾。在這裡的詳述以及例子都 只是示範用，而本發明真正的保護範圍當視後附之申請專 利範圍所界定者爲準。 (請先閱讀背面之注意事項再填寫本頁)Notes and re-installation I The size of the paper is applicable to the Chinese National Standard (CNS) A4 specification (210 X 297 public love) 483747 A7 B7 5 05 8twf. Doc / 008 5. Probability of the invention description ((() 520) or The similarity 値 (step 510) is compared (step 350 in FIG. 3). The probability or similarity 値 is the reference module (step 520) and the at least one calculated from the formatted digital signal. The degree of similarity of the parameters. The calculated similarity 値 provides a correlation measure of the degree of similarity between the at least one parameter and the corresponding reference module. In an exemplary embodiment of the present invention, the data processing The processor 220 uses a Viterbi Decoder to calculate a similar probability (step 510). In an exemplary embodiment, the data processor 220 repeats (step 530) for any one stored in the storage module 230 Calculate a probability or similarity for 36 types of reference modules (step 510), where each of the 36 types of reference modules corresponds to the status of 36 respiratory systems listed in Table 1. Although this embodiment uses wheat Terbi decoder, We can still use other techniques to compare the at least one parameter with the at least one reference module, for example, including dynamic time warping method, vector quantization, or artificial neural network. Artificial neural networks. In the field of speech recognition, "Basics of Speech Recognition", L.Rabiner and B.Juang, 1993, describes Viterbi Decode, dynamic time warping, vector quantization, and artificial Neural network. Referring again to FIG. 3, in one embodiment of the present invention, the data processor 220 uses a training sample of a known condition of the respiratory system to be analyzed, and calculates a representation of the respiratory system 100. A reference module that knows the condition. The training sample is generally composed of sounds generated by a known condition of the respiratory system. In order to form the reference module, the A / D conversion module converts the training sample into digital Signal, and the data processor 22 applies the digital paper size to the Chinese National Standard (CNS) A4 specification (210 X 297 public love) ( Read the notes on the back before filling this page) 1 am— · 1 ϋ i_i I ον ϋ · ϋ ϋ ϋ I ϋ I DESCRIPTION OF THE INVENTION ((1) The signal is formatted to provide a formatted digital signal, and at least one parameter is calculated to form a corresponding reference module for a known condition to be diagnosed (steps 330, 335, 340). As a result, the reference module includes at least one parameter corresponding to a training sample of a known condition retained in the storage module 230 to be diagnosed. For any other known condition to be diagnosed, the data processor 220 uses another training sample for the other known condition to calculate other reference modules representing the known condition of the other respiratory system. In one embodiment of the present invention, the storage device 230 includes at least one reference module calculated by using a Hidden Markov model according to each identified and diagnosed condition. The data processor 220 uses training samples of each respiratory system condition to calculate a hidden Marco module. For example, the data processor 220 uses a formatted digital signal obtained according to the training sample to calculate a reference module in which the condition corresponds to a hidden Marco module. "Basics of Speech Recognition", L. Rabiner and B.Juang, 1993, describes a hidden Marco module used in speech recognition. Referring to FIG. 3, in order to determine whether the formatted digital signal matches the reference module, the data processor 220 determines (step 360) the at least one parameter or a group of parameters corresponding to one of the reference modules, Whether it is within an acceptable range of similarity. In more detail, the data processor 220 makes a decision and a selection from the situation of the respiratory system 100 stored in the storage module 230 and represented by at least one reference module (step 360). As part of the decision and selection of a possible situation, before providing an output, this paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 public love) (Please read the precautions on the back before filling out this page ) [Equipment -------- Order --------- Line. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 483747 A7 B7 5 05 8twf.doc / 0〇8 V. Description of the invention ( A) The data processing benefit 220 only selects, or prefers, the most similar situations, and filters out unacceptable or dissimilar situations. Fig. 6 shows an exemplary flowchart of a procedure 600 that includes steps for selecting a situation in accordance with the present invention. Referring to FIG. 6, in an exemplary embodiment according to the present invention, the data processor 220 uses the steps shown in the program 600 to determine and select a situation (step 360). The data processor 22 uses the received probability from the comparison step (step 35) to evaluate the probability corresponding to each situation (step 61). The data processor 22 has selected the five most probable 値 (step 62). If the maximum five 値 have an acceptable probability (step 63), then the data processor 22 Provide the condition to be output (step 650 '660). If any of the five conditions cannot represent the true condition of the problem, the data processor 220 selects only those acceptable conditions (step 640), and provide an appropriate output (steps 650, 660) to the output module 240. The output may contain detailed medical information for the selected situation (s) (step 66). As mentioned in the flowchart, the output can be displayed to a user by the output module in a convenient mode (step 650). Referring to FIG. 3 again, the output module 240 provides an indication of the status to a user (Step 370). In an exemplary embodiment, the output module includes, for example, a display ^ 42 and a printer 244 to provide the user with detailed medical information corresponding to the condition (step 66). Therefore, the present invention provides new devices and methods for diagnosing the condition of the respiratory system. (Please read the precautions on the back before filling out this page)-Install -------- Order --------- Line j Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs This paper is suitable for China National Standard (CNS) A '! Specifications mo X 297 public love) 483747 5058twf.doc / 008 V. Description of the invention ((u) The preferred embodiment of the present invention has been disclosed as above, the purpose is to show and present its indications. It is not meant to be thorough. To mention or limit the present invention to the precise form of the disclosure, and through the above teaching revelation or implementation of the present invention, modifications and diversification may be made. For example, the present invention is not limited to the disclosure in The device in Fig. 1 is generally implemented by various methods, and the scope of the present invention is defined in the scope of patent application and its equivalent meaning. Anyone skilled in this art will not depart from the spirit and scope of the present invention. It can be used for various changes and retouching. The detailed description and examples here are for demonstration purposes only, and the true scope of protection of the present invention shall be determined by the scope of the attached patent application. (Please read the precautions on the back first (Fill in this page again)
訂---------線I 經濟部智慧財產局員工消費合作社印製 張 紙 本 適 規 4 )A S) N (C 样Order --------- Line I Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs on paper. Compliance 4) A S) N (C sample
97 2 X97 2 X