kNN Sampling for Personalised Human Activity Recognition

Sani, Sadiq; Wiratunga, Nirmalie; Massie, Stewart; Cooper, Kay

doi:10.1007/978-3-319-61030-6_23

Cited by 36 publications

(17 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Different feature representation approaches for HAR have been proposed, from shallow hand-crafted features to frequency transformation features e.g. Fast Fourier Transform (FFT) and Discrete Cosine Transform (DCT) coefficients, and more recently, deep learning approaches [9,17,18]. All these approaches have had some degree of success and setbacks in performance [15].…”

Section: Introductionmentioning

confidence: 99%

Learning Deep and Shallow Features for Human Activity Recognition

Sani

Massie

Wiratunga

et al. 2017

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

selfBACK is an mHealth decision support system used by patients for the self-management of Lower Back Pain. It uses Human Activity Recognition from wearable sensors to monitor user activity in order to measure their adherence to prescribed physical activity plans. Different feature representation approaches have been proposed for Human Activity Recognition, including shallow, such as with hand-crafted time domain features and frequency transformation features; or, more recently, deep with Convolutional Neural Net approaches. The different approaches have produced mixed results in previous work and a clear winner has not been identified. This is especially the case for wrist mounted accelerometer sensors which are more susceptible to random noise compared to data from sensors mounted at other body locations e.g. thigh, waist or lower back. In this paper, we compare 7 different feature representation approaches on accelerometer data collected from both the wrist and the thigh. In particular, we evaluate a Convolutional Neural Net hybrid approach that has been shown to be effective on image retrieval but not previously applied to Human Activity Recognition. Results show the hybrid approach is effective, producing the best results compared to both hand-crafted and frequency domain feature representations by a margin of over 1.4% on the wrist.

show abstract

Section: Introductionmentioning

confidence: 99%

Learning Deep and Shallow Features for Human Activity Recognition

Sani

Massie

Wiratunga

et al. 2017

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…Considering that the benefits of physical activity are dependent on their correct performance, the control of adherence to activity programs at home is of great interest, which can also be evaluated through accelerometry [181]. This is also the purpose of the "selfBACK" m-Health decision support system, which is introduced in [182,183] for the self-control of lumbar pain. This system controls the activity of the subjects through human activity recognition by using accelerometer sensors to evaluate their adherence to the prescribed physical activity plans.…”

Section: Benefits Of Physical Activity To Painmentioning

confidence: 99%

Sensor Technologies to Manage the Physiological Traits of Chronic Pain: A Review

Naranjo-Hernández

Reina‐Tosina

Roa

2020

Sensors

View full text Add to dashboard Cite

Non-oncologic chronic pain is a common high-morbidity impairment worldwide and acknowledged as a condition with significant incidence on quality of life. Pain intensity is largely perceived as a subjective experience, what makes challenging its objective measurement. However, the physiological traces of pain make possible its correlation with vital signs, such as heart rate variability, skin conductance, electromyogram, etc., or health performance metrics derived from daily activity monitoring or facial expressions, which can be acquired with diverse sensor technologies and multisensory approaches. As the assessment and management of pain are essential issues for a wide range of clinical disorders and treatments, this paper reviews different sensor-based approaches applied to the objective evaluation of non-oncological chronic pain. The space of available technologies and resources aimed at pain assessment represent a diversified set of alternatives that can be exploited to address the multidimensional nature of pain.

show abstract

“…If motion patterns between users are different, this follows quite naturally from the assumption behind the classification models that the distribution between training and application must be the same. However, because collecting a sufficient amount of data for each target user is not practical, much research focused on methods to fine-tune general models given only a small amount of labeled target data [12,15,19,32,35,36,38,39]. Acquiring even a small set of labeled data may be costly and impractical.…”

Section: Related Workmentioning

confidence: 99%

Incremental Real-Time Personalization in Human Activity Recognition Using Domain Adaptive Batch Normalization

Mazankiewicz

Böhm

Bergés

2020

Proc. ACM Interact. Mob. Wearable Ubiquitous Technol.

View full text Add to dashboard Cite

Human Activity Recognition (HAR) from devices like smartphone accelerometers is a fundamental problem in ubiquitous computing. Machine learning based recognition models often perform poorly when applied to new users that were not part of the training data. Previous work has addressed this challenge by personalizing general recognition models to the unique motion pattern of a new user in a static batch setting. They require target user data to be available upfront. The more challenging online setting has received less attention. No samples from the target user are available in advance, but they arrive sequentially. Additionally, the motion pattern of users may change over time. Thus, adapting to new and forgetting old information must be traded off. Finally, the target user should not have to do any work to use the recognition system by, say, labeling any activities. Our work addresses all of these challenges by proposing an unsupervised online domain adaptation algorithm. Both classification and personalization happen continuously and incrementally in real time. Our solution works by aligning the feature distributions of all subjects, be they sources or the target, in hidden neural network layers. To this end, we normalize the input of a layer with user-specific mean and variance statistics. During training, these statistics are computed over user-specific batches. In the online phase, they are estimated incrementally for any new target user.

show abstract

kNN Sampling for Personalised Human Activity Recognition

Cited by 36 publications

References 16 publications

Learning Deep and Shallow Features for Human Activity Recognition

Learning Deep and Shallow Features for Human Activity Recognition

Sensor Technologies to Manage the Physiological Traits of Chronic Pain: A Review

Incremental Real-Time Personalization in Human Activity Recognition Using Domain Adaptive Batch Normalization

Contact Info

Product

Resources

About