AccNet24: A deep learning framework for classifying 24-hour activity behaviours from wrist-worn accelerometer data under free-living environments. (April 2023)

• Record Type: Journal Article
• Authors: Farrahi, Vahid; Muhammad, Usman; Rostami, Mehrdad; Oussalah, Mourad
• Is Part Of: International journal of medical informatics. Volume 172(2023)
• Publication Date: 2023-04
• Subjects: Activity recognition -- Transfer learning -- Raw acceleration -- Machine learning -- Activity classification; Medical informatics -- Periodicals; Information science -- Periodicals; Computers -- Periodicals; Medical technology -- Periodicals; Medical Informatics -- Periodicals; Technology, Medical -- Periodicals; Computers; Information science; Medical informatics; Medical technology; Electronic journals; Periodicals; Electronic journals; 610.285
• DOI: 10.1016/j.ijmedinf.2023.105004
• Languages: English
• ISSNs: 1386-5056

Abstract

Highlights: AccNet24 converted acceleration signal into signal images. Deep features were extracted from signal images using transfer learning. BiLSTM was used to classify the 24-hour activity behaviours. AccNet24 achieved > 95 % accuracy on unseen data. The next generation accelerometry analytics may rely on deep learning methodologies. Abstract: Objective: Although machine learning techniques have been repeatedly used for activity prediction from wearable devices, accurate classification of 24-hour activity behaviour categories from accelerometry data remains a challenge. We developed and validated a deep learning-based framework for classifying 24-hour activity behaviours from wrist-worn accelerometers. Methods: Using an openly available dataset with free-living wrist-based raw accelerometry data from 151 participants (aged 18–91 years), we developed a deep learning framework named AccNet24 to classify 24-hour activity behaviours. First, the acceleration signal (x, y, and z-axes) was segmented into 30-second nonoverlapping windows, and signal-to-image conversion was performed for each segment. Deep features were automatically extracted from the signal images using transfer learning and transformed into a lower-dimensional feature space. These transformed features were then employed to classify the activity behaviours as sleep, sedentary behaviour, and light-intensity (LPA) and moderate-to-vigorous physical activity (MVPA) using a bidirectional long short-term memory (BiLSTM) recurrent neural network. AccNet24 was trained and validated with data from 101 and 25 randomly selected participants and tested with the remaining unseen 25 participants. We also extracted 112 hand-crafted time and frequency domain features from 30-second windows and used them as inputs to five commonly used machine learning classifiers, including random forest, support vector machines, artificial neural networks, decision tree, and naïve Bayes to classify the 24-hour activity behaviour categories. Results: Using the same training, validation, and test data and window size, the classification accuracy of AccNet24 outperformed the accuracy of the other five machine learning classification algorithms by 16%–30% on unseen data. Conclusion: AccNet24, relying on signal-to-image conversion, deep feature extraction, and BiLSTM achieved consistently high accuracy (>95 %) in classifying the 24-hour activity behaviour categories as sleep, sedentary, LPA, and MVPA. The next generation accelerometry analytics may rely on deep learning techniques for activity prediction.