Solving the Problem of Unifying the Size of Large Squares of Grid in ECG Images

Abstract

ECG signals are electrical changes associated with heart activity as a function of time. These signals are recorded using an ECG acquisition system on special paper containing a grid divided by large and small squares. In this study, ECG images were processed and noise was eliminated, followed by image segmentation and restoration. In order to describe and represent the image as a one-dimensional signal, a new effective method was proposed to solve the problem of unifying the size of the grid squares in the captured ECG images despite of  the difference in camera distance from the paper tape. It was possible to extract the length of the grid segments in ECG images, whatever the imaging method, camera type or vertical distance from the paper. The grid square of ECG paper was extracted in 97% of the total number of studied images. The proposed system was able to convert the 100% studied images into ECG signals. The resulting signals are of a storage size less than the size of the ECG images, helping to automate the clinics of specialists in heart disease or internal diseases widely. The images were processed and the proposed method was applied using MATLAB.