Attention-Based Convolutional Neural Network for Weakly Labeled Human Activities’ Recognition With Wearable Sensors

Traditional methods of human activity recognition usually require a large amount of strictly labeled data for training classifiers. However, it is hard for one to keep a fixed activity when collecting desired activity data by wearable sensors, and the weakly labeled data inevitably occurs in the process of data collection. For now, human activity recognition methods have seldom been researched according to weakly labeled data, which deserves deep investigation. In this paper, we proposed a novel attention-based human activity recognition method to process the weakly labeled activity data. The traditional convolutional neural network (CNN)-based human activity recognition is modified by attention mechanism, which computes the compatibility between the global features extracted at the final fully connected layers and the local features extracted at a given convolutional layer. The attention-based CNN architecture can amplify the salient activity information and suppress the irrelevant and potentially confusing information by weighing up their compatibility. Our methods are compared with two state-of-the-art methods, CNN and DeepConvLSTM. The experimental results show that our model is comparably well on the traditional UCI HAR dataset and outperforms them on the weakly labeled dataset in accuracy. Our method can greatly facilitate the process of sensor data annotation and makes data collection easier.

I. Introduction

In the recent years, the wide diffusion of mobile devices has made human activity recognition(HAR) based on wearable sensors [1] become a new research point in the field of artificial intelligence and pattern recognition [2], [3], and there are prevailing applications which benefit from HAR include sports activity detection [4], smart homes [5] and health support [6], et al. Those sensors such as accelerometers, gyroscopes and magnetometers [7], which are embedded on mobile devices, can generate time-series data for HAR. Traditional methods, which have been developed to facilitate human activity recognition, are inside the range of supervised learning. For instance, the pervious methods include SVM [8] and Random Forest [9], which require to extract handcrafted features as the inputs of classifiers. Later, deep learning, and in particular, convolutional neural networks, has been diffusely used in the field of HAR. Although deep learning models have excellent performance in HAR, there are some challenges need to be addressed, the main one of which is the need for labeled datasets for ground truth annotation [10].