ABSTRAKSI: Elektrokardiogram (EKG) adalah suatu sinyal hasil rekaman aktifitas listrik otot jantung dengan memasang elektroda pada badan. Rekaman EKG ini digunakan oleh dokter ahli untuk menentukan kondisi jantung dari pasien. Seiring perkembangan telemedicine, kebutuhan pengiriman rekaman EKG ini dibutuhkan dalam kondisi real time, sehingga waktu pengiriman pun diharapkan secepat mungkin. Oleh karena itu, perlu dilakukan kompresi data sinyal EKG agar pengiriman memiliki delay rendah.
Tugas Akhir ini menganalisis algoritma kompresi sinyal EKG berbasis wavelet dengan delay rendah. Algoritma ini mengatur ukuran frame untuk mencapai delay rendah, tetapi tetap mempertahankan kualitas sinyal rekonstruksi. Target analisis adalah delay di bawah satu detik, Compression Ratio di atas dua, Percent Root mean square Difference di bawah 10%, dan Cross Correlation di atas 95%. Untuk mencapai delay rendah dan kualitas tinggi, ini membutuhkan frame size adjustment, wavelet compression, dan run length encoding. Analisis dilakukan pada 3 tipe sinyal, yaitu Normal Sinus Rhytm, Arrhytmia, dan Atrial Fibrillation masing-masing sinyal diwakili 5 sinyal. Panjang masing-masing sinyal berbeda sesuai frekuensi sampling masing-masing.
Dari hasil analisis, didapat kesimpulan bahwa dalam panjang sinyal yang sama, semakin banyak sampel yang diolah dalam satu frame akan membuat delay semakin rendah, tetapi kualitas sinyal rekonstruksi akan menurun. Nilai optimum diambil dari rata-rata nilai setiap parameter yang dihasilkan. Untuk tiga sinyal yang diuji akan mencapai nilai optimum saat framing sebanyak 10% jumlah sampel asli dan Threshold Multiply Factor sebesar 90. Algoritma ini mendapat hasil yang baik pada sinyal Arrhytmia dengan hasil Compression Ratio (CR) 9,62527, Percent Root mean square Difference 0,60602%, Cross Correlation 99,99807%, dan delay 0,225554 detik.
Kata Kunci : delay, frame size adjustment, wavelet compression, run length encoding, Compresion Ratio, Percent Root-mean square Difference, Cross Correlation.ABSTRACT: Electrocardiogram (ECG) is a signal resulted by heart muscle electric activity by put electrodes to the body. ECG recordings were used by the specialist to determine the patient‟s heart condition. Along with the development of telemedicine, the delivery of ECG recordings are needed in real time conditions, so that the delivery times was expected as soon as possible. Therefore, it is necessary for the ECG signal data compression during transmission can be done quickly so it will has a low delay.
This Final Project analyze wavelet based ECG signals compression algortihm with a low delay property. The algorithm set the frame size to achieve a low delay, while maintaining reconstructed signal quality. The analysis targets are getting delay under one seconds, Compression Ratio above two, Percent Root mean square Difference below 10%, and Cross Correlation aove 95%. To attain both low delay and high quality, it employs frame size adjustment, wavelet compression, and run length encoding. Analysis is done to 3 types of signal which are Normal Sinus Rhytm, Arrhytmia, and Atrial Fibrillation each represented by five signal. The length of each types is different depend on their sampling frequency.
From the system test and analysis in summary for same signal length, processing more samples in one frame will lower the delay, but worse the reconstructed signal quality. The optimum value were taken from each means of the resulted parametre. For three tested signal, reach the optimum value when 10% of original samples framing and 90 Threshold Multiply Factor is done. This algorithm got a good result on the Arrhytmia signal which has Compression Ratio (CR) 9,62527, Percent Root mean square Difference (PRD) 0,60602%, Cross Correlation 99,99807%, and delay 0,225554 seconds.
Keyword: delay, frame size adjustment, wavelet compression, run length encoding, Compression Ratio, Percent Root mean square Difference, Cross Correlation.