Effects of sliding window variation in the performance of acceleration-based human activity recognition using deep learning models

Abstract

Deep learning (DL) models are very useful for human activity recognition (HAR); these methods present better accuracy for HAR when compared to traditional, among other advantages. DL learns from unlabeled data and extracts features from raw data, as for the case of time-series acceleration. Sliding windows is a feature extraction technique. When used for preprocessing time-series data, it provides an improvement in accuracy, latency, and cost of processing. The time and cost of preprocessing can be beneficial especially if the window size is small, but how small can this window be to keep good accuracy? The objective of this research was to analyze the performance of four DL models: a simple deep neural network (DNN); a convolutional neural network (CNN); a long short-term memory network (LSTM); and a hybrid model (CNN-LSTM), when variating the sliding window size using fixed overlapped windows to identify an optimal window size for HAR. We compare the effects in two acceleration sources’: wearable inertial measurement unit sensors (IMU) and motion caption systems (MOCAP). Moreover, short sliding windows of sizes 5, 10, 15, 20, and 25 frames to long ones of sizes 50, 75, 100, and 200 frames were compared. The models were fed using raw acceleration data acquired in experimental conditions for three activities: walking, sit-to-stand, and squatting. Results show that the most optimal window is from 20–25 frames (0.20–0.25s) for both sources, providing an accuracy of 99,07% and F1-score of 87,08% in the (CNN-LSTM) using the wearable sensors data, and accuracy of 98,8% and F1-score of 82,80% using MOCAP data; similar accurate results were obtained with the LSTM model. There is almost no difference in accuracy in larger frames (100, 200). However, smaller windows present a decrease in the F1-score. In regard to inference time, data with a sliding window of 20 frames can be preprocessed around 4x (LSTM) and 2x (CNN-LSTM) times faster than data using 100 frames.

A prediction method for influenza during COVID-19 in Shanxi Province, China: LSTM-SARIMA hybrid model based on Singular Spectrum Analysis

“…This may be attributed to the fact that the LSTM model, a nonlinear model that mimics the learning and reasoning process of the human brain, can directly train and predict data, compensating for the information loss caused by the difference method [17] . Furthermore, the sliding window method was used to form a series of overlapping samples, which can effectively utilize the temporal characteristics of the sequence and reduce the time variability caused by abnormal data, which was consistent with the research results of other scholars [18,19] . The influenza sequence in Shanxi Province was affected by the COVID-19, which had complex time characteristics, and the prediction accuracy of a single model was not well.…”

Age-Related Reliability of B-Mode Analysis for Tailored Exosuit Assistance

“…This approach enabled the inference of smaller subsegments of fascicles of arbitrary lengths. Although the sliding window is a very popular technique for training processes on CNN [ 42 , 43 ], the novelty of this research is the combination of the approach to multiple LSTM units for fascicle length extraction. While the proposed model was trained on fixed-length sequences, no a priori information was provided in input to the model on the length of the inputs, suggesting that the model would successfully deal with varying-length inputs and maintain good performances across participants and tasks.…”

Smartphone User Identification/Authentication Using Accelerometer and Gyroscope Data

“…Utilizing the overlapping sliding window technique for preprocessing low-dimensional time series data results in reduced latency, improved Accuracy, and lower processing costs, particularly when employing a smaller window size [54]. In order to improve the Accuracy, reduce the feature extraction time, and give more comprehensive user data so the model can learn from them, each user signal was split into small segments in this study using the overlapping sliding window technique.…”