Executive Summary : | Cardiovascular disease (CVD) is the leading cause of mortality in India, accounting for 24.8% of all deaths. The lack of timely medical intervention and proper medication due to insufficient medical facilities contributes to this issue. A low-cost and user-friendly system is proposed to self-monitor heart sound signals in daily life scenarios, such as speaking, walking, moving, or traveling. The system will use a wireless module using an accelerometer and a wired module using a stethoscope, connected via Bluetooth and 3.5mm jacks on smartphones. The signal will be converted into digital form using an inbuilt analog-to-digital converter. A graphical user interface (GUI) will be designed to assist users in placing the sensor at the correct location. Discrete wavelet transform (DWT) based denoising algorithms will be implemented for noise suppression in daily life scenarios. An automatic quality assessment parameter will select the best subsequence from the recorded signal, and if not medically acceptable, another signal will be recorded. The selected subsequence will be processed for abnormality detection, with features extracted from the obtained DWT coefficient and a computationally efficient machine learning technique used to classify the signal as normal vs abnormal. The GUI will also facilitate better interaction with the system and communication with medical experts. The inclusion of low-cost sensors, inbuilt components, GUI, and low computation cost algorithm will lead to a low-cost, portable, and user-friendly system for frequent heart check-ups. Medical experts will appreciate this system, which will improve accuracy and speed up diagnosis, and reduce the burden of stretched medical facilities, especially during pandemic situations. |