Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

Difference from Background: Limit of Detection01:05

Difference from Background: Limit of Detection

6.9K
The limit of detection (LOD) is the smallest amount of analyte that can be distinguished from the background noise. The LOD value corresponds to the concentration at which the analyte signal is three times larger than the standard deviation of the blank signal. Below this value, the analyte signal cannot be differentiated from the background noise. It is calculated by dividing the calibration slope by 3 times the standard deviation of the blank signals.
The LOD indicates the presence or absence...
6.9K
IR Frequency Region: Fingerprint Region01:03

IR Frequency Region: Fingerprint Region

1.1K
IR spectra are divided into two main regions: the diagnostic region and the fingerprint region. The diagnostic region of the spectrum lies above 1500 cm−1. The absorptions resulting from single-bond vibrations of the N–H, C–H, and O–H stretch at higher wavenumbers and appear on the left side of the spectrum. The stretching absorptions of the C≡C and C≡N occur between 2100–2300 cm−1. In contrast, those arising from stretching absorptions of the...
1.1K

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Discussion on "INTACT: a method for integration of longitudinal physical activity data from multiple sources" by Jingru Zhang, Erjia Cui, Hongzhe Li, and Haochang Shou.

Biometrics·2026
Same author

Bayesian Decision-Making Shapes Phenotypic Landscapes from Differentiation to Cancer.

Entropy (Basel, Switzerland)·2026
Same author

Ownership change in American nursing homes during the COVID-19 pandemic and the relationship with measures in Donabedian's model of care quality.

The Gerontologist·2026
Same author

Ecological and stochastic determinants of the growth and persistence of the oral pathogen Porphyromonas gingivalis.

NPJ systems biology and applications·2026
Same author

Establishing global standards on wearable technology for measuring mobility in ageing populations: an international consensus exercise.

Age and ageing·2026
Same author

Urban energy load dataset of Chattogram: Peak and off-peak variability insights.

Data in brief·2025

Related Experiment Video

Updated: Sep 3, 2025

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

4.0K

Exploring Orientation Invariant Heuristic Features with Variant Window Length of 1D-CNN-LSTM in Human Activity

Arnab Barua1, Daniel Fuller2, Sumayyah Musa2

  • 1Department of Computer Science, Faculty of Science, Memorial University of Newfoundland, St. John's, NL A1C 5S7, Canada.

Biosensors
|July 27, 2022
PubMed
Summary

Optimal window lengths for human activity recognition (HAR) using smartphone accelerometers are explored. Accuracy saturates around 80% with window lengths over 65, with performance varying for stationary versus non-stationary activities.

Keywords:
CNNLSTMaccelerometerhuman activityinter-participant evaluationorientation invariantsmartphoneswindow length

More Related Videos

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment
06:49

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Published on: December 11, 2015

9.0K
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

9.1K

Related Experiment Videos

Last Updated: Sep 3, 2025

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

4.0K
Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment
06:49

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Published on: December 11, 2015

9.0K
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

9.1K

Area of Science:

  • Computer Science
  • Biomedical Engineering
  • Signal Processing

Background:

  • Human Activity Recognition (HAR) commonly employs deep neural networks (DNNs) with accelerometer data.
  • Challenges in HAR include sensor orientation variability and arbitrary selection of input window lengths for DNNs.
  • Existing research often uses Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, or hybrid CNN-LSTM models.

Purpose of the Study:

  • To investigate the impact of varying window lengths on the performance of a 1D-CNN-LSTM model for HAR.
  • To evaluate the effectiveness of orientation-invariant heuristic features in mitigating sensor orientation challenges.
  • To analyze the recognition performance for six distinct human activities under arbitrary smartphone orientation.

Main Methods:

  • Utilized a 1D-CNN-LSTM architecture for human activity recognition.
  • Employed orientation-invariant heuristic features derived from accelerometer data.
  • Tested various window lengths using data from 42 participants performing sitting, lying, walking, and running activities with smartphones in their pockets.

Main Results:

  • Average classification accuracy reached saturation at approximately 80% (± 8.07%) for window lengths exceeding 65 units.
  • The study identified a trade-off: precision, recall, and F1-score for stationary activities (sitting, lying) decreased as window length increased.
  • Conversely, recognition performance for non-stationary activities (walking, running) improved with longer window lengths.

Conclusions:

  • Window length significantly influences HAR performance, with optimal performance achieved beyond a 65-unit window for the tested 1D-CNN-LSTM model.
  • Orientation-invariant features combined with appropriate window length selection can address sensor orientation issues in HAR.
  • The findings suggest distinct optimal window lengths for stationary versus non-stationary activity recognition.