Abstract:

BACKGROUND: Electrocardiogram (ECG) and other recorded physiological signal curves is an important basis for diagnosis of cardiovascular diseases. With the development of Internet technology and the wide applications of HIS, texts, sounds, and images have already become the digital forms.
OBJECTIVE: To integrate existing paper-recorded ECG graphs to the hospital database for convenient query, statistics and information management and promotion of hospital efficiency, an extraction method of ECG curves based on Gaussian blur technique is presented.
METHODS: Gaussian blur and OSTU binary algorithm were used for the separation of the background grid and original ECG curves, followed by some simple post-processing on the results.
RESULTS AND CONCLUSION: This ECG curves extraction method based on Gaussian blur technique can effectively remove the background grid and keep the intact ECG curves during the course of binarization for varieties of ECG curve pictures. It can serve as a good foundation for digitalization of ECG curves.