1D-CNN-LSTM Hybrid-Model-Based Pet Behavior Recognition through Wearable Sensor Data Augmentation
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
The number of healthcare products available for pets has increased in recent times, which has prompted activeresearch into wearable devices for pets. However, the data collected through such devices are limited by outliersand missing values owing to the anomalous and irregular characteristics of pets. Hence, we propose petbehavior recognition based on a hybrid one-dimensional convolutional neural network (CNN) and long shorttermmemory (LSTM) model using pet wearable devices. An Arduino-based pet wearable device was firstfabricated to collect data for behavior recognition, where gyroscope and accelerometer values were collectedusing the device. Then, data augmentation was performed after replacing any missing values and outliers viapreprocessing. At this time, the behaviors were classified into five types. To prevent bias from specific actionsin the data augmentation, the number of datasets was compared and balanced, and CNN-LSTM-based deeplearning was performed. The five subdivided behaviors and overall performance were then evaluated, and theoverall accuracy of behavior recognition was found to be about 88.76%.
Abstract
The number of healthcare products available for pets has increased in recent times, which has prompted activeresearch into wearable devices for pets. However, the data collected through such devices are limited by outliersand missing values owing to the anomalous and irregular characteristics of pets. Hence, we propose petbehavior recognition based on a hybrid one-dimensional convolutional neural network (CNN) and long shorttermmemory (LSTM) model using pet wearable devices. An Arduino-based pet wearable device was firstfabricated to collect data for behavior recognition, where gyroscope and accelerometer values were collectedusing the device. Then, data augmentation was performed after replacing any missing values and outliers viapreprocessing. At this time, the behaviors were classified into five types. To prevent bias from specific actionsin the data augmentation, the number of datasets was compared and balanced, and CNN-LSTM-based deeplearning was performed. The five subdivided behaviors and overall performance were then evaluated, and theoverall accuracy of behavior recognition was found to be about 88.76%.
참고문헌 (20)
- L. Morgan, "Human–dog relationships during the COVID-19 pandemic : booming dog adoption during social isolation", Humanities and Social Sciences Communications, 7 (2020)[DOI]
- Z. Ng, "The new status quo : enhancing access to human–animal interactions to alleviate social isolation & loneliness in the time of COVID-19", Animals, 11(10) (2021)[DOI]
- G. Cicceri, "A deep learning approach for pressure ulcer prevention using wearable computing", Human-centric Computing and Information Sciences, 10 (2020)[DOI]
- H. Alshammari, "Big IoT healthcare data analytics framework based on fog and cloud computing", Journal of Information Processing Systems, 16(6), pp.1238~1249 (2020)[DOI]
- C. Zhu, "Wearable sensor-based behavioral anomaly detection in smart assisted living systems", IEEE Transactions on Automation Science and Engineering, 12(4), pp.1225~1234 (2015)[DOI]
- "Time series data augmentation for deep learning: a survey" (2020)[URL]
- X. Cui, "Data augmentation for deep neural network acoustic modeling", IEEE, 2015, pp.23(9), 1469~1477
- X. Zhao, "Classification of epileptic IEEG signals by CNN and data augmentation", Proceedings of 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp.926~930 (2020)
- J. Cho, "Design of image generation system for DCGAN-based kids' book text", Journal of Information Processing Systems, 16(6), pp.1437~1446 (2020)[DOI]
- C. Shorten, "A survey on image data augmentation for deep learning", Journal of Big Data, 6 (2019)[DOI]
- "The effectiveness of data augmentation in image classification using deep learning" (2017)[URL]
- T. T. Um, "Data augmentation of wearable sensor data for Parkinson’s disease monitoring using convolutional neural networks", Proceedings of the 19th ACM International Conference on Multimodal Interaction, pp.216~220 (2017)
- D. Van Der Linden, "Buddy's wearable is not your buddy : privacy implications of pet wearables", IEEE Security & Privacy, 17(3), pp.28~39 (2019)[DOI]
- H. F. Nweke, "Multi-sensor fusion based on multiple classifier systems for human activity identification", Human-centric Computing and Information Sciences, 9 (2019)[DOI]
- T. Steels, "Badminton activity recognition using accelerometer data", Sensors, 20(17) (2020)[DOI]
- S. A. Khowaja, "CAPHAR : context-aware personalized human activity recognition using associative learning in smart environments", Human-centric Computing and Information Sciences, 10 (2020)[DOI]
- A. R. Javed, "A collaborative healthcare framework for shared healthcare plan with ambient intelligence", Human-centric Computing and Information Sciences, 10 (2020)[DOI]
- Z. Xu, "Improved 1D-CNNs for behavior recognition using wearable sensor network", Computer Communications, 151, pp.165~171 (2020)[DOI]
- S. Mekruksavanich, "Enhanced hand-oriented activity recognition based on smartwatch sensor data using LSTMs", Symmetry, 12(9) (2020)[DOI]
- J. Kim, "Dog behavior recognition based on multimodal data from a camera and wearable device", Applied Sciences, 12(6) (2022)[DOI]