Using Deep Data Augmentation Training to Address Software and Hardware Heterogeneities in Wearable and Smartphone Sensing Devices

Mathur, Akhil; Zhang, Tianlin; Bhattacharya, Sourav; Veličković, Petar; Joffe, Leonid; Lane, Nicholas D.; Kawsar, Fahim; Lió, Píetro

doi:10.1109/ipsn.2018.00048

Cited by 67 publications

(59 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this way, the system is tested with unseen dataset. We also perform leave-one-subject-out (LOSO) cross-validation for reference by leveraging data augmentation [43,44,45,46], which is widely adopted in the deep learning research that requires large dataset for model training. We apply the jittering method to simulate additive sensor noise using random parameters as described in [47] for label-preserving augmentation.…”

Section: Discussionmentioning

confidence: 99%

Sensor Data Acquisition and Multimodal Sensor Fusion for Human Activity Recognition Using Deep Learning

Chung

Lim

Noh

et al. 2019

Sensors

138

View full text Add to dashboard Cite

In this paper, we perform a systematic study about the on-body sensor positioning and data acquisition details for Human Activity Recognition (HAR) systems. We build a testbed that consists of eight body-worn Inertial Measurement Units (IMU) sensors and an Android mobile device for activity data collection. We develop a Long Short-Term Memory (LSTM) network framework to support training of a deep learning model on human activity data, which is acquired in both real-world and controlled environments. From the experiment results, we identify that activity data with sampling rate as low as 10 Hz from four sensors at both sides of wrists, right ankle, and waist is sufficient in recognizing Activities of Daily Living (ADLs) including eating and driving activity. We adopt a two-level ensemble model to combine class-probabilities of multiple sensor modalities, and demonstrate that a classifier-level sensor fusion technique can improve the classification performance. By analyzing the accuracy of each sensor on different types of activity, we elaborate custom weights for multimodal sensor fusion that reflect the characteristic of individual activities.

show abstract

Section: Discussionmentioning

confidence: 99%

Sensor Data Acquisition and Multimodal Sensor Fusion for Human Activity Recognition Using Deep Learning

Chung

Lim

Noh

et al. 2019

Sensors

138

View full text Add to dashboard Cite

show abstract

“…Recurring neural networks (RNNs) remove temporal addiction and slowly acquire information over time to transmit sensory input to understand human behavior. • Deep neural networking can be detachable and scalable into interconnected networks with a global optimization feature that promotes various profound learning strategies like profound communication learning [14], deep active education [15], a framework for deeper attention [16] and other approaches that are not systemic and effective [17], [18]. Works which take these techniques into account serve to numerous deep learning challenges.…”

Section: B Context Of Deep Learningmentioning

confidence: 99%

“…Several deep learning models have been developed to solve heterogeneity problems caused by different sensor instances. Data augmentation with GANs [18] has been a notable work. The growth in data is a compromise for improved training sets to satisfy the need of a powerful profound learning paradigm for both scale and efficiency.…”

Section: Diversitymentioning

confidence: 99%

RETRACTED ARTICLE: Human activity recognition with deep learning: overview, challenges and possibilities

Kumar

Chauhan

2021

CCF Trans. Pervasive Comp. Interact.

View full text Add to dashboard Cite

The growing use of sensor tools and the Internet of Things requires sensors to understand the applications. There are major difficulties in realistic situations, though, that can impact the efficiency of the recognition system. Recently, as the utility of deep learning in many fields has been shown, various deep approaches were researched to tackle the challenges of detection and recognition. We present in this review a sample of specialized deep learning approaches for the identification of sensor based human behavior. Next, we present the multi-modal sensory data and include information for the public databases which can be used in different challenge tasks for study. A new taxonomy is then suggested, to organize deep approaches according to challenges. Deep problems and approaches connected to problems are summarized and evaluated to provide an analysis of the ongoing advancement in science. By the conclusion of this research, we are answering unanswered issues and providing perspectives into the future.

show abstract

“…In our previous work, we quantified the accuracy loss caused by microphone heterogeneity in mobile and embedded microphones [12,13] in the context of a keyword detection model. We recorded a large spoken keyword dataset simultaneously on three embedded microphones (namely Matrix Voice, ReSpeaker, PlugUSB) and trained microphone specific CNN models for performing keyword detection.…”

Section: Microphone Variability In Speech Modelsmentioning

confidence: 99%

Unsupervised domain adaptation for robust sensory systems

Mathur

Isopoussu

Bianchi-Berthouze

et al. 2019

Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of The

Self Cite

View full text Add to dashboard Cite

Despite significant advances in the performance of sensory inference models, their poor robustness to changing environmental conditions and hardware remains a major hurdle for widespread adoption. In this paper, we introduce the concept of unsupervised domain adaptation which is a technique to adapt sensory inference models to new domains only using unlabeled data from the target domain. We present two case-studies to motivate the problem and highlight some of our recent work in this space. Finally, we discuss the core challenges in this space that can trigger further ubicomp research on this topic.

show abstract

Using Deep Data Augmentation Training to Address Software and Hardware Heterogeneities in Wearable and Smartphone Sensing Devices

Cited by 67 publications

References 26 publications

Sensor Data Acquisition and Multimodal Sensor Fusion for Human Activity Recognition Using Deep Learning

Sensor Data Acquisition and Multimodal Sensor Fusion for Human Activity Recognition Using Deep Learning

RETRACTED ARTICLE: Human activity recognition with deep learning: overview, challenges and possibilities

Unsupervised domain adaptation for robust sensory systems

Contact Info

Product

Resources

About