Multimodal features play a key role in wearable sensor based human activity recognition (HAR). Selecting the most salient features adaptively is a promising way to maximize the effectiveness of multimodal sensor data. In this regard, we propose a 'collect fully and select wisely' principle as well as an interpretable parallel recurrent model with convolutional attentions to improve the recognition performance. We first collect modality features and the relations between each pair of features to generate activity frames, and then introduce an attention mechanism to select the most prominent regions from activity frames precisely. The selected frames not only maximize the utilization of valid features but also reduce the number of features to be computed effectively. We further analyze the accuracy and interpretability of the proposed model based on extensive experiments. The results show that our model achieves competitive performance on two benchmarked datasets and works well in real life scenarios.
|Title of host publication||2018 International Joint Conference on Neural Networks, IJCNN 2018 - Proceedings|
|Publication date||10 Oct 2018|
|Publication status||Published - 10 Oct 2018|
|Event||2018 International Joint Conference on Neural Networks, IJCNN 2018 - Rio de Janeiro, Brazil|
Duration: 8 Jul 2018 → 13 Jul 2018
|Conference||2018 International Joint Conference on Neural Networks, IJCNN 2018|
|City||Rio de Janeiro|
|Period||08/07/2018 → 13/07/2018|
|Series||Proceedings of the International Joint Conference on Neural Networks|
Bibliographical notePublisher Copyright:
© 2018 IEEE.
- deep learning
- wearable sensors