Bogazici University Smartphone Accelerometer Sensor Dataset

Mobile devices especially smartphones have gained high popularity and become a part of daily life in recent years. Smartphones have built-in motion sensors such as accelerometer, gyroscope and orientation sensors. Recent researches on smartphones show that behavioral biometrics can be obtained from the smartphone motion sensors. In this context, we develop an Android application that collects accelerometer sensor data while user playing a game. This application records all accelerometer data and touch event information while users touch the screen. We perform two experiments and collect two different data using this application. In the first experiment, we collect data from 107 child users whose age vary from 4 to 11, and 100 adult users whose age are between 16 and 55. This dataset includes more than 11.000 taps data for child and adult users, in total. In the second experiment, data is collected from 60 female and 60 male users aged 17-57 for different activities like sitting and walking. There are more than 6.000 taps data for sitting and walking scenarios separately in the second dataset. We use popular Android smartphones in the experiments and they have all 100 Hz sampling rate. These data can be used for behavioral biometric analyses such as user age group and gender detection, user identification and authentication or tap event detection.近年来，移动设备尤其是智能手机已得到极高普及，成为日常生活的组成部分。智能手机内置加速度计（accelerometer）、陀螺仪（gyroscope）与方向传感器（orientation sensors）等运动传感器。现有针对智能手机的相关研究表明，可通过智能手机的运动传感器获取行为生物特征。在此背景下，我们开发了一款安卓（Android）应用程序，用于在用户游玩游戏时采集加速度传感器数据。该应用会记录用户触摸屏幕过程中的全部加速度数据与触摸事件信息。我们依托该应用开展了两项实验，并采集了两组不同的数据。 第一项实验中，我们从107名年龄介于4至11岁的儿童用户，以及100名年龄处于16至55岁的成年用户处采集数据。本数据集总计包含儿童与成年用户共超11000条触摸点击（tap）数据。 第二项实验则面向60名女性与60名男性用户（年龄跨度为17至57岁），针对坐姿、行走等不同活动场景采集数据。第二组数据集分别针对坐姿与行走场景，各包含超6000条触摸点击数据。 本次实验采用的均为主流安卓智能手机，且所有设备均采用100赫兹的采样率。 上述数据可用于行为生物特征分析相关任务，例如用户年龄组与性别识别、用户身份鉴别与认证，以及触摸点击事件检测等。