Contactless Video-Based Heart Rate Monitoring of a Resting and an Anesthetized Pig

Area of Science:

• Veterinary Medicine
• Biomedical Engineering
• Animal Science

Background:

• Heart rate (HR) is a critical indicator of an animal's health, stress, and well-being.
• Current HR monitoring often requires physical contact, limiting continuous assessment.
• Developing non-invasive methods is crucial for improving livestock welfare and management.

Purpose of the Study:

• To develop and validate an algorithm for extracting pig heart rate (HR) from raw video data.
• To establish a foundation for continuous, contactless monitoring of pig health and welfare.
• To assess the algorithm's performance on both anesthetized and resting pigs.

Main Methods:

• Utilized video recordings of pigs equipped with electrocardiography (ECG) as the gold standard (GS).
• Applied a bandpass filter to raw video data to minimize noise.
• Employed a short-time Fourier transform (STFT) method, optimizing window sizes and functions for HR extraction.
• Investigated the effectiveness of different regions of interest (ROI), focusing on the abdomen.

Main Results:

• The algorithm achieved a mean absolute error (MAE) of 2.33 bpm and root mean square error (RMSE) of 3.09 bpm for anesthetized pigs.
• For resting pigs, the algorithm demonstrated an MAE of 4.69 bpm and RMSE of 6.43 bpm.
• Using only the green video channel and focusing on the abdominal ROI yielded superior results and reduced computational complexity.

Conclusions:

• The developed video-based algorithm shows significant potential for contactless HR measurement in pigs.
• This non-invasive approach can facilitate real-time monitoring of pig health and welfare status.
• The findings are of considerable interest to veterinarians and farmers for improved livestock management.