Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation
Collar-mounted canine activity monitors can use accelerometer data to estimate dog activity levels, step counts, and distance traveled. With recent advances in machine learning and embedded computing, much more nuanced and accurate behavior classification has become possible, giving these affordable...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-05-01
|
| Series: | Animals |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-2615/11/6/1549 |
| _version_ | 1797532675428319232 |
|---|---|
| author | Robert D. Chambers Nathanael C. Yoder Aletha B. Carson Christian Junge David E. Allen Laura M. Prescott Sophie Bradley Garrett Wymore Kevin Lloyd Scott Lyle |
| author_facet | Robert D. Chambers Nathanael C. Yoder Aletha B. Carson Christian Junge David E. Allen Laura M. Prescott Sophie Bradley Garrett Wymore Kevin Lloyd Scott Lyle |
| author_sort | Robert D. Chambers |
| collection | DOAJ |
| description | Collar-mounted canine activity monitors can use accelerometer data to estimate dog activity levels, step counts, and distance traveled. With recent advances in machine learning and embedded computing, much more nuanced and accurate behavior classification has become possible, giving these affordable consumer devices the potential to improve the efficiency and effectiveness of pet healthcare. Here, we describe a novel deep learning algorithm that classifies dog behavior at sub-second resolution using commercial pet activity monitors. We built machine learning training databases from more than 5000 videos of more than 2500 dogs and ran the algorithms in production on more than 11 million days of device data. We then surveyed project participants representing 10,550 dogs, which provided 163,110 event responses to validate real-world detection of eating and drinking behavior. The resultant algorithm displayed a sensitivity and specificity for detecting drinking behavior (0.949 and 0.999, respectively) and eating behavior (0.988, 0.983). We also demonstrated detection of licking (0.772, 0.990), petting (0.305, 0.991), rubbing (0.729, 0.996), scratching (0.870, 0.997), and sniffing (0.610, 0.968). We show that the devices’ position on the collar had no measurable impact on performance. In production, users reported a true positive rate of 95.3% for eating (among 1514 users), and of 94.9% for drinking (among 1491 users). The study demonstrates the accurate detection of important health-related canine behaviors using a collar-mounted accelerometer. We trained and validated our algorithms on a large and realistic training dataset, and we assessed and confirmed accuracy in production via user validation. |
| first_indexed | 2024-03-10T11:03:39Z |
| format | Article |
| id | doaj.art-9b2a5deb49cf445589eab1fc18da7c8a |
| institution | Directory Open Access Journal |
| issn | 2076-2615 |
| language | English |
| last_indexed | 2024-03-10T11:03:39Z |
| publishDate | 2021-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Animals |
| spelling | doaj.art-9b2a5deb49cf445589eab1fc18da7c8a2023-11-21T21:20:21ZengMDPI AGAnimals2076-26152021-05-01116154910.3390/ani11061549Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World ValidationRobert D. Chambers0Nathanael C. Yoder1Aletha B. Carson2Christian Junge3David E. Allen4Laura M. Prescott5Sophie Bradley6Garrett Wymore7Kevin Lloyd8Scott Lyle9Pet Insight Project, Kinship, San Francisco, CA 94103, USAPet Insight Project, Kinship, San Francisco, CA 94103, USAPet Insight Project, Kinship, San Francisco, CA 94103, USAPet Insight Project, Kinship, San Francisco, CA 94103, USAPet Insight Project, Kinship, San Francisco, CA 94103, USAPet Insight Project, Kinship, San Francisco, CA 94103, USAWALTHAM Petcare Science Institute, Melton Mowbray, Leicestershire LE14 4RT, UKPet Insight Project, Kinship, San Francisco, CA 94103, USAPet Insight Project, Kinship, San Francisco, CA 94103, USAPet Insight Project, Kinship, San Francisco, CA 94103, USACollar-mounted canine activity monitors can use accelerometer data to estimate dog activity levels, step counts, and distance traveled. With recent advances in machine learning and embedded computing, much more nuanced and accurate behavior classification has become possible, giving these affordable consumer devices the potential to improve the efficiency and effectiveness of pet healthcare. Here, we describe a novel deep learning algorithm that classifies dog behavior at sub-second resolution using commercial pet activity monitors. We built machine learning training databases from more than 5000 videos of more than 2500 dogs and ran the algorithms in production on more than 11 million days of device data. We then surveyed project participants representing 10,550 dogs, which provided 163,110 event responses to validate real-world detection of eating and drinking behavior. The resultant algorithm displayed a sensitivity and specificity for detecting drinking behavior (0.949 and 0.999, respectively) and eating behavior (0.988, 0.983). We also demonstrated detection of licking (0.772, 0.990), petting (0.305, 0.991), rubbing (0.729, 0.996), scratching (0.870, 0.997), and sniffing (0.610, 0.968). We show that the devices’ position on the collar had no measurable impact on performance. In production, users reported a true positive rate of 95.3% for eating (among 1514 users), and of 94.9% for drinking (among 1491 users). The study demonstrates the accurate detection of important health-related canine behaviors using a collar-mounted accelerometer. We trained and validated our algorithms on a large and realistic training dataset, and we assessed and confirmed accuracy in production via user validation.https://www.mdpi.com/2076-2615/11/6/1549canineaccelerometerdeep learningbehavioractivity monitor |
| spellingShingle | Robert D. Chambers Nathanael C. Yoder Aletha B. Carson Christian Junge David E. Allen Laura M. Prescott Sophie Bradley Garrett Wymore Kevin Lloyd Scott Lyle Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation Animals canine accelerometer deep learning behavior activity monitor |
| title | Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation |
| title_full | Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation |
| title_fullStr | Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation |
| title_full_unstemmed | Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation |
| title_short | Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation |
| title_sort | deep learning classification of canine behavior using a single collar mounted accelerometer real world validation |
| topic | canine accelerometer deep learning behavior activity monitor |
| url | https://www.mdpi.com/2076-2615/11/6/1549 |
| work_keys_str_mv | AT robertdchambers deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT nathanaelcyoder deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT alethabcarson deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT christianjunge deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT davideallen deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT lauramprescott deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT sophiebradley deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT garrettwymore deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT kevinlloyd deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation AT scottlyle deeplearningclassificationofcaninebehaviorusingasinglecollarmountedaccelerometerrealworldvalidation |