201110937 六、發明說明： 【發明所屬之技術領域】 本發明係與檢測辨識器有關，特別是指一種心電 辨識器。 【先前技術】 由於心臟敲胁雜完全治_ 病，且其發病時間 往往只有數分鐘的時間而已，因此需要長期追縱觀察。例如： • 狹心症患者在接受檢查、練的過程中，很少會出現持續胸痛 的狹心症症狀’必須對該患者施以運動負荷試驗直到患者感 到胸痛、呼吸困難、域降低、嚴重心律不整為止，此時，即 _叙式的檢測及分析儀n，爾以及分析出心電圖和 血壓變化來判《/f症狀。然而，由於醫院所使用的檢測及分析儀 器’多屬於固定式儀器或整套檢測平台，對患者而言，除了需 支付高額較檢費狀外，更必須長娜住在狀，因而限制 • 了患者行動範圍與正常的生活，因此，-般以具有非侵入性、 攜帶方便、易取得雜的霍特⑽ter)心電圖記錄器，對患者 進行24小0^心電圖（Electrocardiogram, ECG)紀錄，醫師再透 過離線模式尋找異常的心·訊號，以作最有觀患者病情 的診斷。 然而，一但患者返回居家生活時，雖然使用該霍特(Holter) 心電圖e錄器得以24小時紀錄病患^跳異常的發生時段、以 及異常的心電圖訊號，但是病患除了定期至醫院回診之外， 201110937 醫師並無法從患者的日常生 3 中立即侍到有效的異常心電圖 訊號’以作最有利於病情的診斷。 據此，如何開發出一種 4 a 電圖自動辨識器，其可解決上 述問題’即是本發明急欲解 鮮秀的問4’亦是本發明研發的 動機。 【發明内容】 本發明之目的在於提供—種心電圖自細顧，其主要 可自動偵測異常的心電圖訊號，並藉由無線網路自動傳送 至醫院及記錄，讓病患在不需要額外複雜設備下，能有效 摘取及侧異常心頓訊號，而讓醫師從日常生活中可立 即了解病患病情狀況’以作出最有利於病患病情的診斷。 緣是，為了達成前述目的，依據本發明所提供之一種 心電圖自動辨識器，其包含有：一心電圖取得模組，用以產 生對應於心臟的心電圖訊號，且該心電圖訊號具有一 QRS複合 波；一心電圖訊號特徵萃取模組，其係將該QRS複合波區分為 Q-R區段與R-S區段’並對該Q-R區段與R_s區段取複數個樣 點進行轉換，藉以獲得複數筆轉換資料；一類神經辨識器模 組，其具有一輸入層、一存有複數筆比對資料的隱藏層、一累 加層以及一輸出層，其中該轉換資料經該輸入層而至隱藏層 内’藉以與該隱藏層内的比對資料進行比對，比對後取相同類 型的轉換資料與比對資料進行累加並由輸出層輸出一異常訊 201110937 號； =触，総物__所輸出 屮:吊訊號無線網路模組，其將該類神經辨識器模組的輸 :所輸-異常職以鱗通财錢出至遠端的無線通 訊接收器。 有關本發明為達成上述目的，所_之技術、手段及其他 之力效’⑵舉-較佳可行實施例並配合圖式詳細說明如后。 【實施方式】201110937 VI. Description of the Invention: [Technical Field of the Invention] The present invention relates to a detection recognizer, and more particularly to an electrocardiograph. [Prior Art] Since the heart is completely ill-treated, and the onset time is often only a few minutes, it is necessary to observe for a long time. For example: • Patients with angina have few symptoms of angina with persistent chest pain during the examination and training. 'The patient must be tested for exercise stress until the patient feels chest pain, difficulty breathing, domain reduction, severe heart rate At this time, the detection and analyzer n, and the analysis of the electrocardiogram and blood pressure changes to judge the "/f symptom. However, because the detection and analysis instruments used by hospitals are mostly fixed instruments or a complete set of detection platforms, in addition to the high fees and fees, it is necessary for the patients to stay in shape, thus limiting the patients. The range of action and normal life, therefore, the non-invasive, portable, easy to get Hult (10) ter) ECG recorder, the patient's 24 small 0 ^ Electrocardiogram (ECG) record, the doctor then through Offline mode to find abnormal heart and signal for the diagnosis of the most patient's condition. However, once the patient returned to his home life, although the Holter ECG recorder was used to record the 24-hour abnormality of the patient's abnormality and the abnormal electrocardiogram signal, the patient was given regular visits to the hospital. In addition, the 201110937 physician is unable to immediately deliver a valid abnormal ECG signal from the patient's daily life 3 for the most favorable diagnosis. Accordingly, how to develop a 4 a electrogram automatic identifier which can solve the above problem is that the invention is eager to solve the problem 4' which is also the motivation for the development of the present invention. SUMMARY OF THE INVENTION The object of the present invention is to provide an electrocardiogram self-care, which can automatically detect abnormal ECG signals and automatically transmit them to hospitals and records through a wireless network, so that patients do not need additional complicated equipment. Under the circumstance, the patient can effectively pick up the side and the abnormal heart signal, and let the doctor know the patient's condition immediately from daily life to make the diagnosis that is most beneficial to the patient's condition. In order to achieve the foregoing objective, an electrocardiogram automatic identifier according to the present invention includes: an electrocardiogram acquisition module for generating an electrocardiogram signal corresponding to a heart, and the electrocardiogram signal has a QRS complex; An ECG signal feature extraction module, which divides the QRS complex into a QR segment and an RS segment and converts the plurality of samples of the QR segment and the R_s segment to obtain a plurality of conversion data; A type of neural identifier module having an input layer, a hidden layer storing a plurality of pairs of data, an accumulation layer, and an output layer, wherein the conversion data is passed through the input layer to the hidden layer Compare the comparison data in the hidden layer, compare and then take the same type of conversion data and comparison data to accumulate and output an abnormal signal 201110937 by the output layer; = touch, 総 __ output 屮: hoist signal The wireless network module, which outputs the type of the neural identifier module, transmits the abnormality to the remote wireless communication receiver. The present invention has been made in view of the above-described objects, techniques, and other advantages of the invention. [Embodiment]
首先參閱第1圖，本發明實施例所提供的-種心電圖自動 辨識’其主要係由—心電圖取得模組1G、-心電圖訊號 特徵萃取模組2G…類神經辨識器模組3()、—顯示模組4〇 以及一無線網路模組50所構成，其中： 該。電圖取得模組10,其包含有用以取得。、臟心電圖 Λ號11的電極>;、用以放大該心電圖訊號11的放大器、 用以濾、除已放大的该心賴訊號^雜訊的帶通紐器13、以Referring first to FIG. 1 , an automatic electrocardiogram identification provided by an embodiment of the present invention is mainly performed by an electrocardiogram acquisition module 1G, an electrocardiogram signal feature extraction module 2G, a neurological identifier module 3 ( ), The display module 4A and a wireless network module 50 are configured, wherein: The electrogram acquisition module 10 contains useful information to obtain. , the dirty electrocardiogram, the electrode of the 1111, the amplifier for amplifying the electrocardiogram signal 11, the band-passing device 13 for filtering, and the amplified signal of the heartbeat
及用以制該心電圖訊號！i中QRS複合波ii _s c圓㈣R 峰值的R峰值搜尋漁14，藉以產生對應細臟的心電圖訊 號15，且該心電圖訊號具有- QRS複合波151，該QRS複 。波151為心電圖中最明顯的部份，通常呈現高聳尖銳的高頻 波形。 §亥心電圖訊號特徵萃取模組20，其係以該QRS複合波151 中R波最大峰值為中心’將則Rs複合;皮151㊣分解成二個區 奴’即Q-R區段與區段’並對該q_r區段與R_s區段取複 201110937 數個樣點進行轉換，藉以獲得複數筆轉換資料；而該心電圖訊 號特徵萃取模組20中的轉換函數式為： ΨΡι (χρ W) = cpn'+d ρχρ [t]+fp+gp sin(fijw') 其中參數P表7F Q-R區段與s區段的區段編號，參數t表示 樣點編號’ Xp[t]’t=l，2,3, ···，％，表示每個樣點資料，參數 Np表示Q-R區段與R_s區段取樣點個數，本實施例中q_r區 段與R-S區段分別設定25個樣點，整個哪複合波則共有5〇 個樣點’參數D表示碎形維度(Fractal Dimension, FD)，該碎 維度藉於1及2之間，由區段與R_s區段的樣點資料， 求出轉換函數的餘值映射參數③吻吻^’^^為 各轉換資料所需的參數。 筆：=:識器模組30 ’其具有一輸入層、-存有複數 .男'...—藏層、~累加層以及一輪出層，其中哕韓拖次 行比對，比對細l藉以與該隱藏層内的比對資料進 由輸出層輸出h枓進仃累加並 λ 、㊉訊唬；其中該類神經辨識芎槿紐… 入層至隱藏層的數 ^器模組30的輪 双值違异函數式為： ί=ΐ 2σ^ 4=句，“丄 、中 為^藏層函數’私·為轉換資料，—'Iff為 201110937 料 ’ k=l，2, 3 ···忆 一 —’ ’ ·，Κ’參數k表示比對資料的編號，參數κ . 二比^貝^總個數’ 4為歐氏距離平方值，巧為歐氏距 入層時，㈣數調整錄，#雜齡料&輪入至輪 另夕:Λ轉換資料色·得與該比對資料历相加。 算函數式為轉賴組3G的累加層至輸*層的數值運 ί>* :J為輪出層函數，货广為權重值，孖ft . 隱藏層函數，复腌 4 一將相同類型病症所對應的隱藏層輪出相加， 播=’即所測得該心朗訊號15的QRS複合波151轉# 後的轉換資料鱼嗲抨:轉指 症相同貝，胸；Z 的比對資料相同時輸出相加’若病 1定為1 ’該累加層的運算函數式為： ^,wjkmffk , y=l} 2, 3.....m， 表5^輸出Bp點編號，錄m表示輸出節點總個數t 該顯示模組4〇，复_ 一 所輪出的異常峨，換鋪躲職隐3G的輸出層 、3之，該顯示模組40具有一液晶顯开 赏I’ _該顯示模纟且仙 、、、〇收到該異常信號時，則藉由該液晶_ 不螢幕顯示偵測纟Jr果 ^ 且可顯示病症類型、出現異常生理郭 爾》的次數、發並招:靖& 者。 、的時間、及發出警訊的訊息以提醒配戴 201110937 该無線網路模組50’其包括一無線通訊發射器51以及一 無線通訊接㈣52’該遠端無線通訊餘器51將該類神經辨 識器模組3G_dJ騎輸出的異常減以無線通訊方式輪出 至遠端的無線通訊接收器52,可設定該無線通訊發射器51 與無線通訊接收器52的傳輸頻段、傳輸通道、傳輸速率、 Master/Slave屬性，則可達雙向通訊功能；值得一提的適， 當該無線網路模組50的遠端無線通訊接收器52將其所收到 的異常訊號傳送至終端設備53，再透過網際網路平台將異常 訊號傳輸至醫院54以供診斷。 據此，本發明除了可自動偵測異常的心電圖訊號之 外，並可藉由無線網路自動傳送至醫院及記錄，讓病患在 不需要額外複雜設備下，能有效擷取及偵測異常心電圖訊 號’而讓醫師從日常生活中可立即了解病患病情狀況，以 作出最有利於病患病情的診斷。 綜上所述，此實施例及圖示僅為本發明之較佳實施例而 已，當不能以之限定本發明實施之範圍，即大凡依本發明申 請專利範圍所作之均等變化與修飾，皆應屬本發明專利涵蓋 之範圍内。And used to make the ECG signal! The QR peak of the QRS complex ii _s c circle (4) R peak searches for the fish 14 to generate a correspondingly dirty ECG signal 15, and the ECG signal has a -QRS complex 151, which is complex. Wave 151 is the most visible part of the electrocardiogram and usually presents a sharp, high-frequency waveform. §Hai electrocardiogram signal feature extraction module 20, which is centered on the maximum peak value of the R wave in the QRS complex 151, and then Rs is composited; the skin 151 is decomposed into two zone slaves, ie, a QR segment and a segment' The q_r segment and the R_s segment are converted into a plurality of 201110937 samples to obtain a plurality of conversion data; and the conversion function in the ECG signal feature extraction module 20 is: ΨΡι (χρ W) = cpn' +d ρχρ [t]+fp+gp sin(fijw') where parameter P is the segment number of the 7F QR and s segments, and the parameter t is the sample number 'Xp[t]'t=l,2, 3, ···,%, for each sample data, the parameter Np represents the number of sampling points of the QR segment and the R_s segment. In the embodiment, the q_r segment and the RS segment respectively set 25 samples, which of the whole The composite wave has a total of 5 samples. The parameter D represents the Fractal Dimension (FD). The fragmentation dimension is derived between 1 and 2, and the transfer function is obtained from the sample data of the segment and the R_s segment. The residual value mapping parameter 3 kisses ^'^^ is the parameter required for each conversion data. Pen:=: Recognizer module 30' has an input layer, - has a complex number. Male '... - Tibetan layer, ~ accumulated layer and one round out layer, of which 哕韩拖次行对对对对对对l by means of the comparison data in the hidden layer, the output layer is outputted by the output layer, and the λ and the tenth 唬 are included; wherein the type of the neural layer is added to the hidden layer The round-valued double-difference function is: ί=ΐ 2σ^ 4=sentence, “丄, 中为^层层function' private·for conversion data,—“Iff is 201110937 material' k=l,2, 3 ·· ·Recall one-' ' ·, Κ 'parameter k indicates the number of the comparison data, parameter κ. 2 than ^ Bay ^ total number ' 4 is the Euclidean distance squared value, when the Euclidean distance into the layer, (four) number Adjust the record, #多龄料&round to the round another day: Λ conversion data color···················································· ;* :J is the round-out function, the goods are widely weighted, 孖ft. Hidden layer function, re-salted 4, the hidden layer corresponding to the same type of illness is added, broadcast = 'that is measured QR code of Lucent 15 After the conversion of the 151 rpm# fish 嗲抨: the same finger shell, chest; Z comparison data when the output is added 'If the disease 1 is set to 1 ' The cumulative function of the operation function is: ^, Wjkmffk , y=l} 2, 3.....m, Table 5^ Output Bp point number, record m indicates the total number of output nodes t The display module 4〇, complex _ a round of abnormal 峨, The display module 40 has a liquid crystal display display I' _ the display module and the sensation, 〇, 〇 receives the abnormal signal, then the liquid crystal _ The screen displays the number of symptoms, the number of abnormal physiological genes, and the time: The network module 50' includes a wireless communication transmitter 51 and a wireless communication connection (4) 52'. The remote wireless communication residual device 51 rotates the abnormality of the output of the neural identifier module 3G_dJ by wireless communication to The remote wireless communication receiver 52 can set the transmission frequency band and transmission of the wireless communication transmitter 51 and the wireless communication receiver 52. The channel, the transmission rate, and the Master/Slave attribute can reach the two-way communication function; it is worth mentioning that when the remote wireless communication receiver 52 of the wireless network module 50 transmits the abnormal signal received by the wireless network module 50 to the terminal The device 53 transmits the abnormal signal to the hospital 54 for diagnosis through the internet platform. Accordingly, the present invention can automatically detect the abnormal ECG signal and can be automatically transmitted to the hospital and the record through the wireless network. In order to enable patients to effectively detect and detect abnormal ECG signals without the need for additional complicated equipment, doctors can immediately understand the patient's condition from daily life to make a diagnosis that is most conducive to the patient's condition. In the above, the embodiments and the drawings are only the preferred embodiments of the present invention, and the scope of the present invention is not limited thereto, that is, the equal changes and modifications made by the scope of the patent application of the present invention should be It is within the scope of the patent of the present invention.
201110937 【圖式簡單說明】 第1圖係本發明示意圖。 【主要元件符號說明】 心電圖取得模組10 QRS複合波111 帶通濾波器13 心電圖訊號15 心電圖訊號特徵萃取模組20 類神經辨識器模組30 顯示模組40 無線網路模組50 無線通訊接收器52 醫院54 心臟心電圖訊虎11 放大器12 R峰值搜尋模組14 QRS複合波151 無線通訊發射器51 終端設備53201110937 [Simple description of the drawings] Fig. 1 is a schematic view of the present invention. [Main component symbol description] ECG acquisition module 10 QRS complex wave 111 bandpass filter 13 ECG signal 15 ECG signal feature extraction module 20 Neurological identifier module 30 Display module 40 Wireless network module 50 Wireless communication reception 52 Hospital 54 Heart ECG Tiger 11 Amplifier 12 R Peak Search Module 14 QRS Complex 151 Wireless Communication Transmitter 51 Terminal Equipment 53