Deep, Convolutional, and Recurrent Models for Human Activity Recognition using Wearables

Hammerla, Nils Y.; Halloran, Shane; Ploetz, Thomas

doi:10.48550/arxiv.1604.08880

Cited by 70 publications

(104 citation statements)

References 4 publications

Supporting

Mentioning

102

Contrasting

Order By: Relevance

“…The CondConv systematically performs better than the DeepConvLSTM, improving the performance by 2.28% on average on the OPPORTUNITY dataset. [19] 78.90% Hammerla et al2016 [36] 74.50%…”

Section: A Experiments Results and Performance Comparisonmentioning

confidence: 96%

Real-time Human Activity Recognition Using Conditionally Parametrized Convolutions on Mobile and Wearable Devices

Cheng¹,

Zhang²,

Tang³

et al. 2020

Preprint

View full text Add to dashboard Cite

Recently, deep learning has represented an important research trend in human activity recognition (HAR). In particular, deep convolutional neural networks (CNNs) have achieved state-of-the-art performance on various HAR datasets. For deep learning, improvements in performance have to heavily rely on increasing model size or capacity to scale to larger and larger datasets, which inevitably leads to the increase of operations. A high number of operations in deep leaning increases computational cost and is not suitable for real-time HAR using mobile and wearable sensors. Though shallow learning techniques often are lightweight, they could not achieve good performance. Therefore, deep learning methods that can balance the trade-off between accuracy and computation cost is highly needed, which to our knowledge has seldom been researched. In this paper, we for the first time propose a computation efficient CNN using conditionally parametrized convolution for real-time HAR on mobile and wearable devices. We evaluate the proposed method on four public benchmark HAR datasets consisting of WISDM dataset, PAMAP2 dataset, UNIMIB-SHAR dataset, and OPPORTUNITY dataset, achieving state-of-the-art accuracy without compromising computation cost. Various ablation experiments are performed to show how such a network with large capacity is clearly preferable to baseline while requiring a similar amount of operations. The method can be used as a dropin replacement for the existing deep HAR architectures and easily deployed onto mobile and wearable devices for real-time HAR applications.

show abstract

“…The CondConv systematically performs better than the DeepConvLSTM, improving the performance by 2.28% on average on the OPPORTUNITY dataset. [19] 78.90% Hammerla et al2016 [36] 74.50%…”

Section: A Experiments Results and Performance Comparisonmentioning

confidence: 96%

Real-time Human Activity Recognition Using Conditionally Parametrized Convolutions on Mobile and Wearable Devices

Cheng¹,

Zhang²,

Tang³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…One approach provide medication states only during some specific activities, such as walking or non-walking, [10,9,7,8]; however, this approach is unable to provide a continuous monitoring of the subjects as needed for detection of the duration of different medication states. The other approach, similar to our algorithm, makes a medication state detection independent of the activity type [12,13,14,33,11,34], and hence is able to provide duration in different medication states. Our developed algorithm provided the highest sensitivity (94%) and specificity (85%) compared to the state-of-the-art using only two sensors.…”

Section: Comparison To Other Studiesmentioning

confidence: 99%

“…This is problematic as those methods can identify medication states only during those specific activities and are unable to provide a continuous monitoring of the subjects as needed for detection of clinically important information about duration in different medication states. Second, some approaches trade accuracy for continuous monitoring [11,12,13,14,15], or have to use five to seven sensors at different parts of the body in order to provide acceptable accuracy. As a result, the condition under which the subjects need to use the device is very impractical with too many sensors to wear.…”

Section: Introductionmentioning

confidence: 99%

Wearable-based Mediation State Detection in Individuals with Parkinson's Disease

Hssayeni,

Burack,

et al. 2018

Preprint

View full text Add to dashboard Cite

One of the most prevalent complaints of individuals with mid-stage and advanced Parkinson's disease (PD) is the fluctuating response to their medication (i.e., ON state with maximum benefit from medication and OFF state with no benefit from medication). In order to address these motor fluctuations, the patients go through periodic clinical examination where the treating physician reviews the patients' self-report about duration in different medication states and optimize therapy accordingly. Unfortunately, the patients' self-report can be unreliable and suffer from recall bias. There is a need to a technology-based system that can provide objective measures about the duration in different medication states that can be used by the treating physician to successfully adjust the therapy. In this paper, we developed a medication state detection algorithm to detect medication states using two wearable motion sensors. A series of significant features are extracted from the motion data and used in a classifier that is based on a support vector machine with fuzzy labeling. The developed algorithm is evaluated using a dataset with 19 PD subjects and a total duration of 1,052.24 minutes (17.54 hours). The algorithm resulted in an average classification accuracy of 90.5%, sensitivity of 94.2%, and specificity of 85.4%.

show abstract

“…A neural network which is to be run on a cloud or desktop computer can leverage massive computing resources. Therefore, most of the cloud and desktop-based works focus on optimizing the speed of the search process and the accuracy of neural networks [22,26,1,36,37,30,5,25,30,14,29].…”

Section: Introductionmentioning

confidence: 99%

MemNet: Memory-Efficiency Guided Neural Architecture Search with Augment-Trim learning

Liu¹,

Wu²,

Ma³

et al. 2019

Preprint

View full text Add to dashboard Cite

Recent studies on automatic neural architectures search have demonstrated significant performance, competitive to or even better than hand-crafted neural architectures. However, most of the existing network architecture tend to use residual, parallel structures and concatenation block between shallow and deep features to construct a large network. This requires large amounts of memory for storing both weights and feature maps. This is challenging for mobile and embedded devices since they may not have enough memory to perform inference with the designed large network model. To close this gap, we propose MemNet, an augment-trim learning-based neural network search framework that optimizes not only performance but also memory requirement. Specifically, it employs memory consumption based ranking score which forces an upper bound on memory consumption for navigating the search process. Experiment results show that, as compared to the state-of-the-art efficient designing methods, MemNet can find an architecture which can achieve competitive accuracy and save an average of 24.17% on the total memory needed.

show abstract

Deep, Convolutional, and Recurrent Models for Human Activity Recognition using Wearables

Cited by 70 publications

References 4 publications

Real-time Human Activity Recognition Using Conditionally Parametrized Convolutions on Mobile and Wearable Devices

Real-time Human Activity Recognition Using Conditionally Parametrized Convolutions on Mobile and Wearable Devices

Wearable-based Mediation State Detection in Individuals with Parkinson's Disease

MemNet: Memory-Efficiency Guided Neural Architecture Search with Augment-Trim learning

Contact Info

Product

Resources

About