以主成分分析與模糊推論方法判斷心跳種類

近幾年由於飲食習慣的改變，心血管疾病患者的年齡層逐漸下降，心臟疾病所造成的死亡率越來越高，因此監測心臟疾病的技術扮演了重要的角色。傳統的心電圖檢查步驟繁瑣並且需要醫療人員的專業協助，當病患心臟發生輕微異常時，往往到醫院作詳細檢查時已恢復正常而查不出病因。本研究針對心跳監測設計一套適用於居家照護之心電圖即時檢測系統，使用心電圖感測器搭配無線感測網路技術來檢測使用者的心跳快慢與心律變動性的指標，並將檢測結果儲存到資料庫，本研究也整合雲端系統，以網頁方式呈現監測結果，達到輔助診斷的效果。此外，對於心律不整的心跳檢測，本研究使用美國麻省理工學院提供的MIT-BIH資料庫來分析，將常見的心律不整心跳種類NORM、LBBB、RBBB、VPC、APC及PB，使用主成分分析演算法來選取主要特徵，再將選取後的主要特徵以模糊推論方法設計心律不整偵測模組，依據不同心跳的特徵設計對應的模糊規則庫。為了有效降低系統運算複雜度，在不增加誤判率的原則下，將原本六項心跳特徵刪減至五項，並將相鄰的歸屬函數整合，由原本條模糊規則組合簡化至種組合。實驗結果顯示所提系統能達成正常心跳的脈搏數檢測，心律不整心跳 NORM、LBBB、RBBB、VPC、APC與PB之判別準確率分別為97.5%、87.5%、92.5%、100%、95%與100%。實驗結果驗證本研究所提之系統適用於心律不整心跳種類判斷。

關鍵字

模糊推論；無線感測網路；主成分分析；心律不整

並列摘要

For the past few years due to the changes in diet habits, patients with cardiovascular disease become progressively younger, and the rate of deaths caused by heart disease rises; therefore, using new technologies to monitor heart disease play an important role. Traditional ECG inspection procedures can be very complicated and need professional assistance. When minor abnormality occurs the patient’s heart often returns to normal before he/she takes a detailed examination. Thus, the cause of abnormality remains unsolved. In this study, we proposed a suitable design of heartbeat monitoring ECG real-time detection system for home care, which uses the ECG sensors and a wireless sensor network technology to detect the user's heartbeat rates and their variations. This study also integrates with the cloud system, showing the monitoring results on web pages as diagnosing assistance. In addition, the Massachusetts Institute of Technology MIT-BIH database is used to analyze arrhythmia, which is used to select main features from common types of arrhythmia, namely, NORM, LBBB, RBBB, VPC, APC and PB. Based on the selected features an arrhythmia detection module is devised to detect the arrhythmia that is determined by the corresponding fuzzy rules. In order to reduce the system complexity without increasing the false positive rate, we cut heartbeat characteristics from six to five. Besides, the fuzzy rules are simplified from 66 to rules. The experimental results show that the proposed system is able to detect normal and arrhythmia heartbeat detection. For NORM, LBBB, RBBB, VPC, APC and PB the discriminative accuracy rates are 97.5%, 87.5%, 92.5%, 100%, 95% and 100%, respectively. The experimental result shows that the proposed system is suitable for distinguishing the types of arrhythmia heartbeat.

並列關鍵字

Fuzzy inference ； wireless sensor network ； principal component analysis ； arrhythmia

參考文獻

[34]中華民國醫師公會全國聯合會 – 心律變動性分析， http://www.tma.tw/ltk/98520603.pdf

[36]Y. C. Yeh and H. J. Lin, “An analysis of ECG for determining heartbeat case by using the principal component analysis and fuzzy logic,” in Proc. of the 19th National Conf. on Fuzzy Theory and its Applications, Yunlin, Taiwan, pp.18-19, Nov. 2011.

[2]Chin-Feng Lai, Yueh-Min Huang, Jong Hyuk Park and Han-Chieh Chao, “Adaptive Body Posture Analysis for Elderly-Falling Detection with Multisensors,” IEEE Intelligent Systems, vol. 25, no. 2, pp.70-81, March-April 2010.

[3]Paolo Bonato, “Wearable Sensors and Systems,” IEEE Engineering in Medicine and Biology Society, vol.29, no. 3, pp.25-36,May-June 2010.

[5]M.S. Cheng, S.S. Chang, J. Moore, G.D. Luca, S.H. Nawab, S.H. Roy and C.J.D. Luca, “A combined sEMG and accelerometer system for monitoring functional activity in stroke,” IEEE Engineering in Medicine and Biology Society, vol. 17, no. 6, pp.585-594, Dec. 2009.

國際替代計量

以主成分分析與模糊推論方法判斷心跳種類

未授權

主題瀏覽