Human Activity Recognition

FALL-UP dataset

The aim of this project is to assist elderly people using novel human activity recognition machine learning methods.

Description

Layout

Layout of the sensors in the experiments

Photo of the experiment setup

Photo of the experiment setup from a different point of view

COMPONENTS

Subjects

Data

This dataset is publicly available, and it comprises 11 activities performed by 17 subjects during 3 attempts each.

The organization of the dataset is as follows:

  • 17-folders, one per subject:
    • 11-subfolders, one per activity:
      • 3 - subfoldes, one per trial:
        • 1-CSV file containing the sensor signals of the attempt.

Each CSV file contains the following information:

  • [Column 1] Timestamp
  • [Columns 2 - 4] IMU (ankle): accelerometer X, Y & Z
  • [Columns 5 - 7] IMU (ankle): angular velocity X, Y & Z
  • [Column 8] IMU (ankle): luminosity
  • [Columns 9 - 11] IMU (Right pocket): accelerometer X, Y & Z
  • [Columns 12 - 14] IMU (Right pocket): angular velocity X, Y & Z
  • [Column 15] IMU (Right pocket): luminosity
  • [Columns 16 - 18] IMU (belt): accelerometer X, Y & Z
  • [Columns 19 - 21] IMU (belt): angular velocity X, Y & Z
  • [Column 22] IMU (belt): luminosity
  • [Columns 23 - 25] IMU (neck): accelerometer X, Y & Z
  • [Columns 26 - 28] IMU (neck): angular velocity X, Y & Z
  • [Column 29] IMU (neck): luminosity
  • [Columns 30 - 32] IMU (wrist): accelerometer X, Y & Z
  • [Columns 33 - 35] IMU (wrist): angular velocity X, Y & Z
  • [Column 36] IMU (wrist): luminosity
  • [Column 37] Raw value from EEG NeuroSky helmet
  • [Column 38] Infrared Sensor 1
  • [Column 39] Infrared Sensor 2
  • [Column 40] Infrared Sensor 3
  • [Column 41] Infrared Sensor 4
  • [Column 42] Infrared Sensor 5
  • [Column 43] Infrared Sensor 6

If you use this data set, please cite as follows:

  • Lourdes Martínez-Villaseñor, Hiram Ponce, Jorge Brieva, Ernesto Moya-Albor, José Núñez-Martínez, Carlos Peñafort-Asturiano, “UP-Fall Detection Dataset: A Multimodal Approach”, Sensors 19(9), 1988: 2019, doi:10.3390/s19091988.

Downloads

In this section, you can use the following layout to chose the elements you want to download, by selecting a subject, activity and trial. In this layout you can look for specific files, however, if your interest lies in the complete data sets, you can use the links under the "Complete Downloads" heading to get those (unfortunately we don't have a single file for all the images).

Each file contains:

  • Data Set: all acquired data (synchronized, and organized with tags).
  • FeaturesX&Y: features obtained from the collected data. The X value refers to the separation (in seconds) in which the features were obtained and the Y value to the window-length (also in seconds) each feature considered.
  • CameraX: refers to a zip file with the images corresponding to each trial, taken with Camera X.
  • CameraX_OF: refers to a zip file with Camera X's features (Optical Flow).
  • Camera Resized OF: refers to a csv file with both camera's OF, resized to a 20x20 matrix.
  • Camera OF FeaturesX&Y: refers to a csv file with both camera's OF, resized to a 20x20 matrix. From these files the mean was taken as the only feature, where the X value refers to the separation (in seconds) in which the feature was obtained and the Y value to the window-length (also in seconds) the feature considered.

Missing Data

The following data was lost during the trial due to unknown errors within the equipment. Trials 2 and 3 for subject 8 in activity 11 are unavailable

Preliminar FALL-UP dataset

This dataset is publicly available, and it comprises 11 activities, attempted 3 times each by 4 subjects.

The organization of the dataset is as follows:

  • 4-folders, one per subject:
    • 11-subfolders, one per activity:
      • 3-CSV files containing the raw sensor signals, one per attempt

Each CSV file contains the following information:

  • [Column 1] Timestamp
  • [Column 2] Raw value from EEG NeuroSky helmet
  • [Columns 3 - 5] IMU (belt): accelerometer X, Y & Z
  • [Columns 6 - 8] IMU (belt): angular velocity X, Y & Z
  • [Column 9] IMU (belt): luminosity
  • [Columns 10 - 12] IMU (neck): accelerometer X, Y & Z
  • [Columns 13 - 15] IMU (neck): angular velocity X, Y & Z
  • [Column 16] IMU (neck): luminosity
  • [Columns 17 - 19] IMU (pocket): accelerometer X, Y & Z
  • [Columns 20 - 22] IMU (pocket): angular velocity X, Y & Z
  • [Column 23] IMU (pocket): luminosity
  • [Columns 24 - 26] IMU (wrist): accelerometer X, Y & Z
  • [Columns 27 - 29] IMU (wrist): angular velocity X, Y & Z
  • [Column 30] IMU (wrist): luminosity
  • [Column 31] Infrared Sensor 1
  • [Column 32] Infrared Sensor 2
  • [Column 33] Infrared Sensor 3
  • [Column 34] Infrared Sensor 4

The dataset can be downloaded here: Preliminar-Fall-UP Dataset [ZIP-file].