Video based heart rate detection in unrestrained laboratory rats: a feasibility analysis

Abstract

Abstract Continuous monitoring of vital signs in laboratory animals is often essential for reliable scientific results and severity assessment. It still depends on invasive approaches such as transponder implementation, which affect the animals well-being. To minimize this impact, a camera-based method is proposed for heart rate detection by analyzing skin color variations in unrestrained rats. Pulse signals were extracted from video recordings taken with a smart home cage designed in a preceding study and processed using classical signal processing methods. For the successfully detected heart rates, we achieved a mean absolute error (MAE) of 10.6 beats per minute (bpm) and a root mean square error (RMSE) of 13.8 bpm . The results suggest that heart rate detection is feasible. Despite challenges such as difficult lighting conditions and small regions of interest, this study demonstrates the potential for non-invasive heart rate monitoring in freely moving laboratory animals.