Decoding Fatigue: Analyzing Offline Handwriting with Machine Learning to Detect Perceived Exhaustion

Abstract

The quality and readability of an individual’s handwriting and drawing can be influenced by various factors, including their level of physical exertion. This enables us to explore the quantification of exertion by observing an individual’s handwriting. To test this hypothesis, we collected data from 17 participants, building a database of handwriting and drawing samples and their corresponding Borg 10 exertion ratings at the time of drawing. In this paper, we investigate using machine learning techniques to estimate perceived exertion before, during, and after physical activity based on handwriting and drawings. We apply a regression model to compare different drawing tasks and demonstrate that perceived exertion can be predicted using simple line drawings. However, more complex sketches and handwriting demand further research. Our findings suggest that interactive systems could use handwriting and drawing to intervene when users experience excessive discomfort. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s). MUM ’24, December 01–04, 2024, Stockholm, Sweden © 2024 Copyright held by the owner/author(s). ACM ISBN 979-8-4007-1283-8/24/12 https://doi.org/10.1145/3701571.3703393