Year: 2015
Type: dissertation
Abstract: The thesis is concerned with objectively measuring human physical activity through \naccelerometry, and compares the effectiveness of algorithms between obese and non-obese \ngroups. The thesis comprises three studies: \nClassification of Aerobic and Gym-based Exercises from Accelerometer Output. This study \ninvestigated whether accurate classification could be achieved from hip- or ankle-mounted \naccelerometers for a programme of aerobic exercises and free-living activities. It also \nexamined whether accuracy was affected by obesity, and whether a single classifier could be \napplied across BMI groups. The study achieved high classification accuracies (85% for hip \nand 94% for ankle) for both obese and normal BMI groups using the same approach across \ngroups. \nWalking Speed Estimation Using Accelerometry. This study aimed to develop a speed \nestimation model that was applicable across BMI groups, and which utilised a hip-mounted \naccelerometer. To achieve this, multiple accelerometer signal features were evaluated for use \nin a linear speed estimation model, and performance was compared between obese and \nnormal BMI groups. The speed estimation algorithm achieved overall RMSE of 0.08ms-1 \nfor \na mixed BMI group, which is comparable with previous research using homogeneous groups. \nPrediction of Energy Expenditure from Accelerometer Output. This study aimed to identify \nphysiological and anthropometric parameters for use in an improved energy expenditure \nestimation model. Model performance was tested on a mixed BMI group. The energy \nexpenditure prediction model incorporating subject attributes showed around 20% \nimprovement over the standard model. \nThis research found that current approaches to activity classification using accelerometry are \nequally applicable to obese groups and normal BMI groups. Walking speed prediction was \nshown to be possible from a hip-mounted accelerometer for both obese and normal BMI \ngroups. Energy expenditure estimation is improved by including subject-specific parameters \nin the prediction model. Accelerometry is, therefore, a suitable tool for measuring different \naspects of physical activity for obese and mixed BMI groups.
Author: SP Gerrard-Longworth
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1821663136