Integrating the Exercise and Environmental Data into a Digital ECG Structure by Watermarking Technique

Abstract

Exercise test is worldwide recognized as a valuable tool for investigating ST segment-based ischemia markers. Due to load-related risk, the test is reserved for office use, what causes inconvenience, limits patients participation rate and precludes screening for early ischemia stages. Transferring the diagnosis to patients' premises and using everyday activities as a stimulus is an interesting alternative, but needs reliable recording of physical load data. This paper presents a method for integrating the exercise and environmental data into a digital ECG structure by watermarking technique. The method analyses the time-scale ECG representation, detects the bandgap, where the bandwidth of actual cardiac content is lower than the throughput of digital series, detects the noise and replaces it by exercise-related data. Unless in irregular signals, the capacity of data container can accommodate an accompanying accelerometer and environment-related signals without deteriorating the ECG content. This makes possible to perform ECG exercise test in home conditions without additional transmission channels or data structures. The method was tested with CSE database accordingly to EN60601-2-25:2015 and proved the watermarked ECG to maintain the wave borders accuracy within tolerance limits. Consequently, restoration of original ECG record is not necessary. The method was also tested with anonymized stress-test records, which were watermarked with accelerometer data and re-interpreted to yield results fairly comparable to original diagnoses.