{"projects":[{"html":"\n

MIMS-unit algorithm

\n\n

$\"CRAN\nStatus\"$ \n $\"Pkgdown\"$ \n $\".github/workflows/test.yml\"$ \n $\"AppVeyor$ \n $\"\"$

\n\n

Please create github\nissues if you have\nany question related to the package. Please click the following button\nto subscribe to any update on the package (Very important to receive\nnotifications about bugs and bug fixes).

\n\n

Watch

Short introduction

MIMS-unit is abbreviated for Monitor Independent Movement Summary\nunit. This measurement is developed to harmonize the processing of\naccelerometer data from different devices. You may refer to the\nmanuscript for the detail\ndescription of the algorithm.

Copyright and citation

The copyright of the work belongs to Northeastern University, mHealth\nResearch Group. Please kindly cite the\nmanuscript if you\nhave used the package or referred to the algorithm in your work.

Shiny Demo App

You may try to compute MIMS-unit values using our shiny demo app\nhttps://qutang.shinyapps.io/MIMSunit/. Note that the upload file size\nlimit is 50 MB. The usage quote for the server is limited, so we do not\nguarantee the web app is always available to you.

Datasets

All datasets used in the manuscript are available at\nhttps://mhealthgroup.github.io/MIMSunit/articles/datasets.html.

System Requirements

R (>= 3.6.0)
memory (> 4GB)

For Windows

Rtools >=3.5 (see: https://cran.r-project.org/bin/windows/Rtools/)

For Linux (use ubuntu as an example)

Install dependency system packages for devtools: build-essential,\nlibcurl4-gnutls-dev, libxml2-dev, libssl-dev, libcurl4-ssl-dev.

Installation

Stable version

CRAN

install.packages('MIMSunit')\n

Development version

Github

install.packages("devtools")\ndevtools::install_github("mhealthgroup/MIMSunit")\n

Note: It is recommended to use Rstudio when installing the package,\nbecause devtools has some compatible issues with R command line\ninterface.

Usage

MIMSunit::mims_unit(input_dataframe, dynamic_range=c(-3,3), epoch='1 min')\n

Assume the input dataframe is in following format, with the first column\n(timestamp) in POSXlct objects and the device used to collect this\ndata has dynamic range being -3g to 3g. You may set the epoch length to\nbe 1 min, 1 sec, 5 sec, 10 sec and so on.

HEADER_TIME_STAMP,X,Y,Z\n2016-10-03 14:51:14.236,0.007,-0.005,0.984\n2016-10-03 14:51:14.256,0.008,-0.007,0.981\n2016-10-03 14:51:14.276,0.009,-0.006,0.978\n2016-10-03 14:51:14.297,0.009,-0.007,0.984\n2016-10-03 14:51:14.317,0.010,-0.010,0.982\n2016-10-03 14:51:14.337,0.011,-0.010,0.982\n

\n","src":"https://github.com/mHealthGroup/MIMSunit","name":{"en":"MIMSunit: An open-source algorithm to compute movement summary from accelerometer","cn":"MIMSunit: 一种基于加速度计的计算运动总量的开源算法"},"owner":"mHealthGroup","repo":"MIMSunit","license":"Other","official":"https://mhealthgroup.github.io/MIMSunit/","desc":{"en":"MIMS-unit is abbreviated for Monitor Independent Movement Summary unit. This measurement is developed to harmonize the processing of accelerometer data from different devices.","cn":"MIMS-unit指Monitor Independent Movement Summary unit. 这种算法旨在提供一种跨设备的加速度计数据处理和归总算法。"},"tags":{"en":["R","Open source","Algorithm","Mobile health"],"cn":["R","开源","算法","移动健康"]},"page":"1"},{"html":"

Multi-site sensing for activity recognition using accelerometers

This repo hosts the source codes and dataset for Posture and Physical Activity Detection: Impact of Number of Sensors and Feature Type.

Citation

If you have used the codes or referred to the manuscript in your publication, please kindly cite the following paper.

\n
Tang, Q., John, D., Thapa-Chhetry, B., Arguello, D.J. and Intille, S., 2020. Posture and Physical Activity Detection: Impact of Number of Sensors and Feature Type. Medicine & Science in Sports & Exercise. Preprint.
\n

Dependencies

Python >= 3.7
poetry dependency management for python. Install using pip install poetry.
git
graphviz (optional, required to generate workflow diagram pdf)

Get started

Install dependencies

At the root of the project folder, run

> poetry install\n

Reproduce publication results

Run with multi-core processing and memory on a new session folder. Overwrite data or results if they are found to exist.

>> poetry run reproduce --parallel --force-fresh-data=True\n

You may find intermediate and publication results in ./muss_data/DerivedCrossParticipants/product_run.

Sample reproduction results

Check here for a sample reproduction results.

Get help

Run above command at the root of the repository to see the usage of the reproduction script.

>> poetry run python reproduce.py --help\n

The reproduct script will do the following,

Download and unzip the raw dataset
Compute features and convert annotations to class labels
Run LOSO cross validations on datasets of different combinations of sensor placements
Compute metrics from the outputs of the LOSO cross validations
Generate publication tables and graphs

Train and save a model using muss dataset

Make sure you run the reproduce script at first.

You may want to train and save an activity recognition model using the dataset shipped with this repo. To do it, at the root of the project folder, run,

>> poetry run python train_custom_model.py\n

By default, it will train two dual-sensor models (DW + DA) using motion + orientation feature set, each for postures and daily activities.

Get help

>> poetry run python train_custom_model.py --help\n

You may config the type of feature set, the target class labels and the sensor placements used.

For developers

Use your custom model for training and testing

You may check out the two files train_custom_model.py and run_custom_model.py and see how different models are called and used.

\n","src":"https://github.com/qutang/MUSS","name":{"en":"Posture and Physical Activity Detection: Impact of Number of Sensors and Feature Type","cn":"人体姿态和运动监测：不同传感器数目位置以及不同特征的影响研究。"},"owner":"qutang","repo":"MUSS","license":"MIT License","official":null,"desc":{"en":"In this project, we evaluated the effect of single-site versus multisite motion sensing at seven body locations (both ankles, wrists, hips, and dominant thigh) on the detection of physical behavior recognition using a machine learning algorithm.","cn":"本研究旨在定量分析单一或者多个可穿戴传感器对于监测人体姿态和运动的不同影响。"},"tags":{"en":["Python","Open source","Algorithm","Dataset","Mobile health","Machine learning"],"cn":["Python","开源","算法","数据集","移动健康","机器学习"]},"page":"1"},{"html":"

$\"png\"$

Signaligner Pro

What is Signaligner Pro

Signaligner-Pro is an interactive tool for algorithm-assisted exploration and annotation of raw accelerometer data. The tool can be used by researchers using raw accelerometer data to support research in activity recognition/machine learning, exercise science, and sleep quality research among others.

Download Signaligner Pro

v2.3.7-beta

Note that macOS and Linux updates will be released soon for v2.3.7-beta

Research paper and citation request

If you have used Signaligner Pro as part of your research and want to get more details, please read and cite the following paper.

Aditya Ponnada, Seth Cooper, Qu Tang, Binod Thapa-Chhetry, Josh Aaron Miller, Dinesh John, and Stephen Intille. (2021). Signaligner Pro: A Tool to Explore and Annotate Multi-day Raw Accelerometer Data. In 2021 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops). DOI:https://doi.org/10.1109/PerComWorkshops51409.2021.9431110

The original idea for Signaligner Pro stems from the following paper. Please get in touch for more details. The open access manuscript can be found here.

Aditya Ponnada, Seth Cooper, Binod Thapa-Chhetry, Josh Aaron Miller, Dinesh John, and Stephen Intille. (2019). Designing Videogames to Crowdsource Accelerometer Data Annotation for Activity Recognition Research. In Proceedings of the Annual Symposium on Computer-Human Interaction in Play (CHI PLAY ’19), 135–147. DOI:https://doi.org/10.1145/3311350.3347153

$\"gif\"$

Starting up

Select the appropriate operating system and download the package
Unzip the package
If you are using Windows OS, click on the Windows Batch file to run the software
if you are using Mac OS, run the app
This starts Signaligner Pro’s launcher interface

In-built test data

The tool comes packaged with test data to run
On launcher, click on “Open Test Dataset”
It will launch the interface in your default browser
The test data is a simulated sinusoidal data along the X, Y, Z access

Example dataset

An example dataset is available here to test the tool. For convinience, we have included only two hours worth of raw data collected at 100Hz. Please get in touch if you want to try the tool with a larger dataset (7-days or more).

Import your own data

\n
The tool supports data from research-grade Actigraph (GT3X+ or link) as well as with raw data collected from commercial smartwatches and fitness trackers.
\n
\n
The data should be in the following format (standard export format for Actigraph devices).
\n

------------ Data File Created By ActiGraph GT3X+ ActiLife v6.13.1 Firmware v1.3.0 date format M/d/yyyy at 100 Hz  Filter Normal -----------,,\nSerial Number: XXXXXXXXX,,\nStart Time 22:00:00,,\nStart Date 1/26/2017,,\nEpoch Period (hh:mm:ss) 00:00:00,,\nDownload Time 17:31:48,,\nDownload Date 1/01/1901,,\nCurrent Memory Address: 0,,\nCurrent Battery Voltage: 3.89     Mode = 12,,\n--------------------------------------------------,,\nAccelerometer X,Accelerometer Y,Accelerometer Z\n0,0,0\n-0.543,-0.199,0.504\n-0.367,-0.188,-0.414\n-0.273,-0.219,-0.813\n-0.27,-0.223,-0.949\n-0.301,-0.223,-0.969\n

The first ten lines contain header information. Your imported dataset must have at least the Start Time, Start Date, and Sampling Rate (e.g., 100 Hz) information to process and prepare the dataset for visualization and annotation. Pleae get in touch for questions related to converting your sensor data format into Signaligner Pro format.

To import your own dataset, click on “Import Single Dataset”.
If you have data from multiple participants, please select “Import Multiple Datasets”.
You will also be asked to confirm if your participant dataset has data from multiple sensors. For instance, you may have dataset where each participant wore sensors on the wrist and ankle.
If you have data from multiple sensors, please make sure that you place files from different sensors (but for the same participant) in the same sub-folder of the participant.
Once imported, Signaligner Pro will process your data for visualization. This may take several minutes, depending on the size of the dataset.

Labeling data with algorithms

Signaligner Pro is pre-packaged with three algorithms. NOTE: These algorithms are only compatible with the data from Actigraph devices. Please get in touch to run these algorithms for other sensor models.
MUSS: Classifies data into 7 physical activities namely (to be updated). More details can be found here (TBA)
SWAN: Classifies data into sleep, wear, and non-wear states. More details can be found here (TBA)
QC: Checks if there are segments of bad data in the dataset that you may want to ignore. More details can be found here (TBA)
For your imported dataset, select an algorithm of your choice to pre-label the data with. This process may take several minutes

Importing your own labels

You can import your own label files or labels created from other algorithms on the imported dataset. Please make sure that the start and end times of the labels lie within the start and end times of the the imported dataset. The data can be imported in the following format:

"START_TIME","STOP_TIME","PREDICTION","SOURCE","LABELSET"\n"2019-06-19 16:10:06.400","2019-06-19 23:49:53.600","Wear","Player","DEFAULT"\n"2019-06-19 23:49:53.600","2019-06-20 06:36:43.200","Sleep","Player","DEFAULT"\n"2019-06-20 06:36:43.200","2019-06-21 00:32:40.000","Wear","Player","DEFAULT"\n

PREDICTION is the actual label or classification by the algorithm. If the label source is an algorithm, we recommend using “Algo”, or else use “Player”. LABELSET is used to identify the session labels. Please provide a different name if there already exists a default labelset.

Configuring labels

The tool allows you to configure label colors and text
Please go to /users/Documents/SignalignerData/Datasets/(Dataset folder name)/config.json, as shown in the example below.

{\n    "title": "ADITYA_2017_06_02_RAW",\n    "tile_size": 1024,\n    "tile_subsample": 4,\n    "zoom_max": 7,\n    "length": 47485200,\n    "start_time_ms": 1495823940000,\n    "sample_rate": 80,\n    "start_day_idx": 1,\n    "range_min": -8,\n    "range_max": 8,\n    "range_unit": "g",\n    "sensors": [\n        { "sname": "ADITYA_2017_06_02_RAW", "color": [ 1.00, 1.00, 1.00 ], "channels": [ "X", "Y", "Z" ] }\n    ],\n    "channels": [\n        { "cname": "X", "color": [ 0.80, 0.20, 0.20 ] },\n        { "cname": "Y", "color": [ 0.20, 0.80, 0.20 ] },\n        { "cname": "Z", "color": [ 0.20, 0.20, 0.80 ] }\n    ],\n    "labels": [\n        { "lname": "Wear", "color": [ 0.00, 0.90, 0.30 ] },\n        { "lname": "Nonwear", "color": [ 0.00, 0.50, 0.50 ] },\n        { "lname": "Sleep", "color": [ 0.00, 0.00, 0.80 ] },\n        { "lname": "Walking", "color": [ 0.20, 0.00, 0.00 ] },\n        { "lname": "Running", "color": [ 0.70, 0.00, 0.00 ] },\n        { "lname": "Standing", "color": [ 0.50, 0.50, 0.00 ] },\n        { "lname": "Biking", "color": [ 0.80, 0.50, 0.60 ] },\n        { "lname": "Sitting", "color": [ 0.20, 0.00, 0.50 ] },\n        { "lname": "Brisk Walking", "color": [ 0.65, 0.30, 0.00 ] }\n    ]\n}\n

This configuration file can be edited to change the g-range, measurement unit, label colors, as well as sensor channels (e.g., when using a diaxial accelerometer instead of a triaxial accelerometer).

Importing your own video

The tool supports video importing as to display snapshots under your data. Video importing can be done through the launcher. Supported video types are mp4 and mov.

To import a video, first you must to supply a data sensor (.csv file) to be imported with.
In the launcher, click on “Import Single Datasets” or “Import Multiple Datasets”, and import a .mp4 or .mov file along with at least 1 .csv file.
If you are asked to confirm if your dataset has data from multiple sensors, select “Yes.”
Once imported, Signaligner Pro will process your video and data for visualization.
Then you can view your data by clicking “Open Selected Dataset.”

Notes

Signaligner represents time and sample intervals as half-open intervals that include the begin time/sample but not the end time/sample, i.e. [begin, end).

Questions?

For questions related to repo access, feature contribution, data processing scripts, sample datasets, and interface code, please contact:\nAditya Ponnada and Prof. Seth Cooper

For questions related to MUSS and/or QC algorithm, please contact:\nQu Tang

For questions related to SWAN and/or QC algorithm, please contact:\nBinod Thapa-Chhetry

For collaboration-related questions, please contact:\nProf. Stephen Intille, Prof. Seth Cooper, or Prof. Dinesh John

Acknowledgements

The software development was supported by the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health (NIH) under award number UH2EB024407. The work was also supported by NU-TECH AWS credits award from Northeastern University, Boston, MA. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

This is a joint effort between Northeastern University’s mHealth Research Group, Crowdgames Lab, and Exercise Physiology Lab. We are also sincerely thankful to our teammates from sandboxnu.com for their support in improving this software as we continue to work with them.

\n","src":"https://github.com/crowdgames/signaligner-web","name":{"en":"Signaligner Pro: A professional mobile sensor data visualization, annotation tool","cn":"Signaligner Pro: 一种专业的移动传感器数据可视化以及标注工具"},"owner":"crowdgames","repo":"signaligner-web","license":"License not set","official":"https://signaligner.org","desc":{"en":"Signaligner Pro is an interactive tool for algorithm-assisted exploration and annotation of raw accelerometer data. The tool can be used by researchers using raw accelerometer data to support research in activity recognition/machine learning, exercise science, and sleep quality research among others.","cn":"Signaligner Pro是一种可视化交互工具。它支持算法辅助的数据探索以及数据标注。该工具能够帮助研究人员直观地探索高频可穿戴传感器数据，相比其他工具，它有优良的交互性和速度，能够支持交互探索长时间巨量的传感器数据。"},"tags":{"en":["JavaScript","Python","Software","Mobile health"],"cn":["JavaScript","Python","软件","移动健康"]},"page":"1"},{"html":"

Moving Towards a Real-Time System for Automatically Recognizing Stereotypical Motor Movements in Individuals on the Autism Spectrum Using Wireless Accelerometry

This paper extends previous work automatically detecting stereotypical motor movements (SMM) in individuals on the autism spectrum. Using three-axis accelerometer data obtained through wearable wireless sensors, we compare recognition results for two different classifiers – Support Vector Machine and Decision Tree – in combination with different feature sets based on time-frequency characteristics of accelerometer data. We use data collected from six individuals on the autism spectrum who participated in two different studies conducted three years apart in classroom settings, and observe an average accuracy across all participants over time ranging from 81.2% (TPR: 0.91; FPR: 0.21) to 99.1% (TPR: 0.99; FPR: 0.01) for all combinations of classifiers and feature sets. We also provide analyses of kinematic parameters associated with observed movements in an attempt to explain classifier-feature specific performance. Based on our results, we conclude that real-time, person-dependent, adaptive algorithms are needed in order to accurately and consistently measure SMM automatically in individuals on the autism spectrum over time in real-word settings.

Citation

If you have used the datasets or source codes from the repository, please kindly cite the following publication.

\n
Goodwin, M.S., Haghighi, M., Tang, Q., Akcakaya, M., Erdogmus, D. and Intille, S., 2014, September. Moving towards a real-time system for automatically recognizing stereotypical motor movements in individuals on the autism spectrum using wireless accelerometry. In Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing (pp. 861-872).
\n

More details

Please refer to Instruction on Google Doc for a detail description about how to use the dataset and replicate the results in the paper

\n","src":"https://github.com/qutang/stereotypy_ubicomp_14","name":{"en":"Detect stereotypical motor movements in individuals on the autism spectrum using wireless accelerometry","cn":"使用无线可穿戴加速度计来监测自闭症患者的刻板化动作"},"owner":"qutang","repo":"stereotypy_ubicomp_14","license":"MIT License","official":"https://docs.google.com/document/d/12cjQ6QPVeTjPgOZZtoWGJ0Wqh9KEk20LOLi3qEW17D4/edit#","desc":{"en":"This paper describes an algorithm that automatically detects stereotypical motor movements (SMM) in individuals on the autism spectrum using three-axis accelerometer data obtained through wearable wireless sensors.","cn":"本项目旨在开发一种基于可穿戴加速度计的机器学习算法来自动监测自闭症患者的刻板化动作。"},"tags":{"en":["R","MATLAB","Open source","Algorithm","Dataset","Mobile health","Machine learning"],"cn":["R","MATLAB","开源","算法","数据集","移动健康","机器学习"]},"page":"1"},{"html":"

Automated Detection of Puffing and Smoking with Wrist Accelerometers

This repository provides detailed descriptions of the RealSmoking dataset in addition to step-by-step instructions for replicating results presented in Tang’s paper published in the proceedings of PervasiveHealth 14.

Abstract

Real-time, automatic detection of smoking behavior could lead to novel measurement tools for smoking research and “just-in-time” interventions that may help people quit, reducing preventable deaths. This paper discusses the use of machine learning with wrist accelerometer data for automatic puffing and smoking detection. A two-layer smoking detection model is proposed that incorporates both low-level time domain features and high-level smoking topography such as inter-puff intervals and puff frequency to detect puffing then smoking. On a pilot dataset of 6 individuals observed for 11.8 total hours in real-life settings performing complex tasks while smoking, the model obtains a cross validation F1-score of 0.70 for puffing detection and 0.79 for smoking detection over all participants, and a mean F1-score of 0.75 for puffing detection with user-specific training data. Unresolved challenges that must still be addressed in this activity detection domain are discussed.

Citation

Please cite the paper in any presentations or publications if using this dataset or source codes.

\n
Tang, Q., Vidrine, D., Crowder, E., and Intille, S. 2014. Automated Detection of Puffing and Smoking with Wrist Accelerometers. 8th International Conference on Pervasive Computing Technologies for Healthcare, ICST (2014).
\n

Repo structure

* Root Path\n* data\n    * original\n    * featureset (very large size ~ 3.0G)\n    * pkl\n* src\n* results\n* supplements\n    * publication_plots\n    * consolidate_plots\n    * posture_example_plots\n    * puffing_example_plots\n

\n
data/original contains raw sensor data and annotation files. These data may be downloaded here.
\n
\n
data/featureset contains all generated feature sets used by the detection model. These data may be downloaded here.
\n
Note that this folder will not normally be included into the version control repository, as they are intermediate results and can be very large size.
\n
\n
data/pkl contains all binary intermediate files that can be loaded with the python serialization package pickle. These data may be downloaded here.
\n
Note that these intermediate files are featureset files in binary format, with 50% less size). They are not commited to the version control system.
\n
\n
src contains the python source code used in the paper.
\n
\n
results stores all results generated by the source code in .csv format.
\n
\n
supplements stores all plots.
\n
- publication_plots stores all plots used in the final publication.
- consolidate_plots stores supplimentary plots showing and comparison of model prediction and visual inspection along with annotations.
- posture_example_plots stores raw sensor data plot along with annotations for different posture examples.
- puffing_example_plots stores raw sensor data plot along with annotations for different puffing examples.
\n

Original dataset structure

Each session represents a single subject. There are in total seven sessions, please refer to Table 2 in Tang’s paper for the statistical details of each session. And refer to data/statistics_dataset for the original results.

Inside data/original folder you will find a list of .csv files, all the files have similar filename pattern, used to distinguish different sessions, sensor location and dataset type.

[session name]_[sensor location].[dataset_type].csv\n

For example, file session1_DAK.raw.csv represents the raw 3-axis accelerometer data file for session 1 on dominant ankle. And file session2.annotation.csv represents the annotation file for session 2.

Note that there is only one annotation file for each session so sensor_location is omitted.

There are four raw accelerometer data files for each session at four sensor locations,

dominant wrist (DW)
dominant ankle (DAK)
nondominant wrist (NDW)
dominant arm (DAR)

Note that for the experiments run in the paper, only wrist accelerometers are used, please refer to the paper for detailed explanation.

Raw sensor data format

The accelerometer we used is Actigraph GT3X, which has a dynamic range of ±4g and stores value using 10 bits, it’s a 3-axis linear accelerometer. It stores in csv format, with the first column being the unix timestamp, and column 2 to 4 for x, y and z axis.

1319714602650,0.15625,-0.4375,-0.109375\n

There is no header row. Data has already been converted to -4g to +4g from voltage.

Annotation data format

Annotation file has standard csv format with header describes the meaning of each column.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

Column name	Meaning
STARTTIME	start time of current annotation
ENDTIME	end time of current annotation
COLOR	unused
posture	annotation for body posture, could be walking, sitting, standing, lying
activity	annotation for subject activity
smoking	annotation for smoking status, could be either smoking or not-smoking
puffing	annotation for puffing status, could be no-puff, left-puff or right-puff
puff index	the index (start from 0) of current puff, if two annotation has the same puff index, that means they actually belong to the same puff. If it’s a no-puff annotation, the index will be -1
prototypical?	A binary value (0 or 1) indicates whether this puff is prototypical puff. If it’s no-puff, the value will be -1
potential error?	A binary value indicates whether this annotation might be wrong due to human error
note	additional comments on this annotation
link	filename of the corresponding puff example plot in the `puffing_example_plots` folder

Feature set data format

Inside the data/featureset or data/pkl folder, you will find the feature set data files that may also be reproduced by running the codes. Files in data/featureset folder are in .csv format and files in data/pkl folder are in binary format .pkl. Both will share the same filename and have the same content/values once loaded in the program.

[session name]_[feature set type].[window size].[data type].csv\n[session name]_[feature set type].[window size].[data type].pkl\n

There will be two data types: one is for feature vector (.data.csv or .data.pkl) and the other is for class label (.class.csv or .class.pkl). They are associated with each other.

For example, session1_BFW.40.data.csv represents feature vectors for session 1 and the features are computed with data from both wrists and with window size of 40 samples (which is 1s under 40 Hz sampling rate).

As another example, session_all_DW.600.class.csv represents class labels for all sessions and the corresponding .data.csv is computed with dominant wrist accelerometer data with window size 600 samples (which is 15s under 40 Hz sampling rate).

Feature data format

There will be a header row. The first column is the index of current segment (start from 0). And the rest columns are different features. It’s easy to understand the meaning through the names, or you may refer to the documentation of the source codes.

Class label format

There will be a header row with all necessary or unnecessary columns, you can choose to use any of them.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

Column name	Meaning
segment	index of current segment (start from 0)
STARTTIME	start time of current segment
ENDTIME	end time of current segment
seg duration	the duration of current segment in seconds
session	the number indicate which session current segment belongs to
sensor	which sensor does current segment row belongs to (DW, NDW, and BFW), refer to footnote 5 for explanation.
inside puff/puff duration	The duration of puffing happens within current segment / the duration of this current puffing
inside puff/segment duration	The duration of puffing happens within current segment / the duration of this current segment
prototypical?	Whether it’s prototypical puff or not (0, 1) if no-puff, leave empty string
puff index	the index of puff for current segment from current sensor
puff side	left or right
car percentage	The percentage of time that in-car activity occupies current segment
talk percentage	The percentage of time that talking activity occupies current segment
unknown percentage	The percentage of time that unknown-activity occupies current segment
drinking percentage	The percentage of time that drinking-beverage activity occupies current segment
reading percentage	The percentage of time that reading-paper activity occupies current segment
eating percentage	The percentage of time that eating-a-meal activity occupies current segment
smoking percentage	The percentage of time that smoking activity occupies current segment
computer percentage	The percentage of time that using-computer activity occupies current segment
phone percentage	The percentage of time that using-phone activity occupies current segment
walking percentage	The percentage of time that walking posture occupies current segment
superposition percentage	The percentage of time that activity superposition happens during current segment
name	The name of the class assigned to this segment
target	The index of the class assigned to this segment

Reproduce the results in the publication

Coming soon…

Demo: the complexity of smoking behavior in real world

These videos are shot to show the real world cases of smoking behavior. Videos and corresponding accelerometer signal are put side by side to give a direct impression of how movement could affect and intervene the underlying signal.

Separable concurrent activities

In this video, tester was asked to perform smoking while walking. As you can see from the video, signal was shown to contain additive components from both hand movement (puffing) and body movement (walking). These two components are independent and additive because they are conducted by different body components. This gives us some inspiration when dealing with concurrent activities.

$\"Smoking_demo_1\"$
\nClick image to view the video

Ambiguous hand gestures

In this video, tester was asked to perform puffing, eating and drinking during smoking in a natural way. The signal contains several ups and downs but none of them shows distinguish characteristics only for puffing. In fact, these activites all belong to hand-to-mouse gestures. The differences between these activities are quite minor from the view of the signal. This exposes one of the chanllenges in activity recognition, which is to classify similar movements.

$\"Smoking_demo_2\"$
\nClick image to view the video

A comprehensive episode

This video shows a comprehensive episode of natural smoking behavior including a series of complex activities ongoing at the same time. Signals, as shown on the right side, appear to be quite intervened and complex and lack of visible distinguish characteristics for each type of activities. The activites are changing relatively fast in time, thus makes them even more difficult to be captured and recognized in real time.

$\"Smoking_demo_3\"$
\nClick image to view the video

Contact us

If you have any question about the dataset or source codes, please create a github issue.

\n","src":"https://github.com/qutang/tang_pervhealth_14","name":{"en":"Automated Detection of Puffing and Smoking with Wrist Accelerometers","cn":"使用可穿戴加速度计来监测吸烟动作和行为"},"owner":"qutang","repo":"tang_pervhealth_14","license":"MIT License","official":null,"desc":{"en":"This project aims to detect puffing and smoking behavior in a real-world real-time or near-real-time setting with single or multiple on-body accelerometers.","cn":"本项目旨在开发一种实时的基于可穿戴加速度计的机器学习算法来监测吸烟动作和行为。"},"tags":{"en":["Python","Open source","Algorithm","Dataset","Mobile health","Machine learning"],"cn":["Python","开源","算法","数据集","移动健康","机器学习"]},"page":"1"},{"html":"

$\"poster\"$

TOWER AIRDROP

Tower airdrop is an Android experimental exergame for kids to motivate and keep them sweater. It incorporates simple motion primitives such as moving/shaking, tilting and touch gestures into different game tasks: triggering parachute, avoiding obstacles, and stacking a tower. The Game paces are designed to fit the focus attention of kids. Game tasks are designed to be varying throughout the play to prevent kids getting tired of or losing focus when repeating the movement for a long time.

Currently I’m interested about expanding this project to collect data from players (especially from kids), analyze to see their feedback about game pace, and use the results to improve game pace design: designing an Ecological Momentary Assessment (EMA) based scheme for pop up dialogs and also optimize the difficulty and level up mechanism.

$\"screenshots\"$

Downloads

APK package download

v1.0

Licenses of used images

Box image: LGPL, original designer: Everaldo Coelho
Cloud: CC-BY-NC, original designer: Zhe Zhang
Trophy: Free for non-commercial use
Lightning: Public Domain
Locks: Free vector
Halloween monsters: CC-BY-NC-ND. Exception of limitation: fair use. Original designer: Deleket
Green vampire monster: GNU/GPL

\n","src":"https://github.com/qutang/tower-airdrop","name":{"en":"Tower Airdrop: An Android Experimental Exergame for Kids to Motivate and Keep Them Sweater","cn":"Tower Airdrop: 一个基于安卓的儿童实验性运动体感游戏"},"owner":"qutang","repo":"tower-airdrop","license":"MIT License","official":null,"desc":{"en":"The game incorporates simple motion primitives such as moving/shaking, tilting and touch gestures into different game tasks: triggering parachute, avoiding obstacles, and stacking a tower. The Game paces are designed to fit the focus attention of kids.","cn":"这个实验性游戏将人类的基础动作（比如移动、晃动、倾斜）融入不同的游戏任务中，使用智能手机自带的传感器来感应相应动作。该实验项目旨在通过游戏的方式促进儿童的运动意愿。"},"tags":{"en":["Android","Open source","Game","Mobile health"],"cn":["安卓","开源","游戏","移动健康"]},"page":"1"}],"totalPages":1}

undefined

undefined

undefined

undefined

undefined

undefined

undefined