The Free-living Food Intake Cycle (FreeFIC) Dataset

Important update: Since 6/Jan/2021 the FreeFIC dataset has migrated to Zenodo and can be downloaded by following this link. The dataset has been expanded by merging it with FreeFIC held-out part (additional information can be found in the “A Data Driven End-to-end Approach for In-the-wild Monitoring of Eating Behavior Using Smartwatches” article by Kyritsis et al.), reaching a total of 12 unique subjects (up from 6) and 22 in-the-wild sessions (up from 16). The Zenodo link also provides access to helpful scripts that help extract statistics and visualize the recordings. Finally, from now on we will only support the pickle format.


The Free-living Food Intake Cycle (FreeFIC) dataset was created by the Multimedia Understanding Group towards the investigation of in-the-wild eating behavior. This is achieved by recording the subjects’ meals as a small part part of their everyday life, unscripted, activities. The FreeFIC dataset contains the 3D acceleration and orientation velocity signals (6 DoF) from 16 in-the-wild sessions provided by 6 unique subjects. All sessions were recorded using a commercial smartwatch (Ticwatchâ„¢) while the participants performed their everyday activities. In addition, FreeFIC also contains the start and end moments of each meal session as reported by the participants.

Description

FreeFIC includes 16 in-the-wild sessions that belong to 6 unique subjects. Participants were instructed to wear the smartwatch to the hand of their preference well ahead before any meal and continue to wear it throughout the day until the battery is depleted. In addition, we followed a self-report labeling model, meaning that the ground truth is provided from the participant by documenting the start and end moments of their meals to the best of their abilities as well as the hand they wear the smartwatch on. The total duration of the 16 recordings sums up to 77.32 hours, with a mean duration of 4.8 hours.

FreeFIC is also tightly related to FIC, a dataset we created in order to investigate the in-meal eating behavior. More information about FIC can be found here.

Data Statistics

In-the-wild sessions and Meals

In-the-wild sessionsMeals
#1617
Mean (sec)17,3981,148
Std (sec)4,884502
Median (sec)16,4891,065
Total (sec)278,37819,520
Total (hours)77.325.42

Publications

If you plan to use the FreeFIC dataset or any of the resources found in this page, please cite our work:

@article{kyritsis2020data,   
    title={A Data Driven End-to-end Approach for In-the-wild Monitoring of Eating Behavior Using Smartwatches},   
    author={Kyritsis, Konstantinos and Diou, Christos and Delopoulos, Anastasios},   
    journal={IEEE Journal of Biomedical and Health Informatics}, 
    year={2020},   
    publisher={IEEE} 
}
@inproceedings{kyritsis2017automated, 
    title={Detecting Meals In the Wild Using the Inertial Data of a Typical Smartwatch}, 
    author={Kyritsis, Konstantinos and Diou, Christos and Delopoulos, Anastasios}, 
    booktitle={2019 41th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)},  
    year={2019}, 
    organization={IEEE}
} 

Download

We provide the FreeFIC dataset  in three formats. Use the following links to download the format that best fits your needs:

  • Pickle
  • Mat file
  • H5

Important update: Since 6/Jan/2021 the FreeFIC dataset has migrated to Zenodo and can be downloaded by following this link. The dataset has been expanded by merging it with FreeFIC held-out part (additional information can be found in the “A Data Driven End-to-end Approach for In-the-wild Monitoring of Eating Behavior Using Smartwatches” article by Kyritsis et al.), reaching a total of 12 unique subjects (up from 6) and 22 in-the-wild sessions (up from 16). The Zenodo link also provides access to helpful scripts that help extract statistics and visualize the recordings. Finally, from now on we will only support the pickle format.

Technical details

After loading the dataset with your preferred framework you will find 4 keys:

  • subject_id
  • session_id
  • signals
  • meal_gt

Each of these 4 keys contains a list of length 16 (equal to the number of in-the-wild sessions). Elements across lists are aligned, e.g. the 3rd element of the session_id list corresponds to the 3rd element of the signals list.

subject_id list

Each element of this list corresponds to the unique integer identifier of the subject [1,2,3,4,13,14].

The subject identifier in FreeFIC is in line with the subject identifier in the FIC dataset; i.e. FIC’s subject with subject_id = 2 is the same person as FreeFIC’s subject with subject_id = 2.

session_id list

Each element of this list corresponds to the unique integer identifier of the session [1-5]. It should be noted that not all subjects have the same number of sessions.

signals list

Each element of this list is an M*7 data matrix that contains the timestamps, 3D accelerometer and 3D gyroscope measurements for each meal. Specifically, the first column contains the timestamps in seconds, the second, third and forth columns contain the x,y and z accelerometer values in g and the fifth, sixth and seventh columns contain the x,y and z gyroscope values in degrees/second.

All data are recorded at 100Hz. In addition to that, all sensor streams are transformed in such a way that reflects all participants wearing the smartwatch at the same hand with the same orientation, thusly achieving data uniformity. This transformation is in par with the signals in the FIC dataset.

meal_gt list

Each element of this list is a K*2 matrix. Each row represents the meal intervals for the specific in-the-wild session. The first column contains the timestamps of the meal start moments whereas the second one the timestamps of the meal end moments. All timestamps are in seconds.

Dataset history

  • FreeFIC migrated to Zenodo (6/1/2021).

Contact

For any inquiries about the FIC dataset contact Konstantinos Kyritsis at kokirits@mug.ee.auth.gr