User Age Group Recognition on Smartphones and Tablets Using Gesture Swiping Features

Abstract

Swiping is a common touchscreen interaction method. This study investigated the possibility of recognizing user-age groups automatically from swiping behaviors to support the progression of self-adaptive interfaces and authentication. The dataset was collected from 42 participants of younger adults (20–39 years) and older adults (60+ years). Four directions were performed by each participant (down, up, left, and right) on either a smartphone or minitablet, leading to over 2600 trials. Six features were extracted from the data: force pressure (FP), movement time (MT), swipe count (Swipe No), average distance (Avg Distance), speed, and ratio of MT to FP (RMF). KNN and Euclidean distance (ED) algorithms were applied using three training ratios. Classification accuracy was higher on smartphones than mini-tablets. Notably, younger adults were classified with 100% accuracy on smartphones, while older adults reached 96% accuracy on mini-tablets. Across both devices, younger adults were classified with higher accuracy. MT, Avg Distance, and FP emerged as the most age-sensitive features, whereby MT was highly significant (p < 0.001). The findings indicate the feasibility of swiping gestures to be leveraged for age group classification, supporting the development of novel authentication strategies.