Development of an EMBLE Electronic Stethoscope for Disease Diagnosis of Internal Animal Organs Using a Neural Network

Animal auscultation is one of the primary internal organ disease diagnostic methods in veterinary medicine. The method consists of listening to, and detecting, the physiological and pathological sounds of internal organ function. Using neural network algorithms helps the veterinarian in disease diagnostics and minimizes the risk of error. This technology will be especially helpful for those veterinarians who do not have enough practical experience in diagnosing equine diseases. This article describes a method for creating an electronic stethoscope for animals as part of the primary collection of information for further neural network analysis. The device records audio files to an SD card in WAV format. Audio is recorded in stereo mode with a sampling rate of 44.1 kHz and a depth of 16 bits; such a recording quality is necessary for a neural network. Furthermore, the recorded files will be processed by a neural network to identify pathologies. The device uses an ESP32 microcontroller, and the software is written in C++. To digitize the sound, MEMS microphones are used, working via a I2S interface. In order for the veterinarian to control the data collection process, a headphone jack is provided into which sound is continuously broadcast. The device's operability is confirmed by the wave diagram of auscultation of the lungs and heart obtained using the developed device. The stethoscope under development is portable, with a battery life of up to 5 hours, and allows diagnoses for up to 15 animals per battery charge. This electronic stethoscope is developed by EMBLE LLC company and is a part of the intellectual predictive animal disease diagnostics system.