Methods Inf Med 2016; 55(06): 533-544
DOI: 10.3414/ME15-01-0130
Original Articles
Schattauer GmbH

Automatic Identification of Physical Activity Intensity and Modality from the Fusion of Accelerometry and Heart Rate Data

Fernando García-García
1   Bioengineering and Telemedicine Group, Universidad Politécnica de Madrid, Madrid, Spain
2   Networking Research Center on Bioengineering, Biomaterials and Nanomedicine (CIBER-BBN), Madrid, Spain
,
Pedro J. Benito
3   Laboratory of Exercise Physiology, Facultad de Ciencias de la Actividad Física y del Deporte (INEF), Universidad Politécnica de Madrid, Madrid, Spain
,
María E. Hernando
1   Bioengineering and Telemedicine Group, Universidad Politécnica de Madrid, Madrid, Spain
2   Networking Research Center on Bioengineering, Biomaterials and Nanomedicine (CIBER-BBN), Madrid, Spain
› Author Affiliations
Further Information

Publication History

received: 29 September 2015

accepted: 28 April 2016

Publication Date:
08 January 2018 (online)

Summary

Background: Physical activity (PA) is essential to prevent and to treat a variety of chronic diseases. The automated detection and quantification of PA over time empowers lifestyle interventions, facilitating reliable exercise tracking and data-driven counseling.

Methods: We propose and compare various combinations of machine learning (ML) schemes for the automatic classification of PA from multi-modal data, simultaneously captured by a biaxial accelerometer and a heart rate (HR) monitor. Intensity levels (low / moderate / vigorous) were recognized, as well as for vigorous exercise, its modality (sustained aerobic / resistance / mixed). In to -tal, 178.63 h of data about PA intensity (65.55 % low / 18.96 % moderate / 15.49 % vigorous) and 17.00 h about modality were collected in two experiments: one in free- living conditions, another in a fitness center under controlled protocols. The structure used for automatic classification comprised: a) definition of 42 time-domain signal features, b) dimensionality reduction, c) data clustering, and d) temporal filtering to exploit time redundancy by means of a Hidden Markov Model (HMM). Four dimensionality reduction techniques and four clustering algorithms were studied. In order to cope with class imbalance in the dataset, a custom performance metric was defined to aggregate recognition accuracy, precision and recall.

Results: The best scheme, which comprised a projection through Linear Discriminant Ana -lysis (LDA) and k-means clustering, was evaluated in leave-one-subject-out cross-validation; notably outperforming the standard industry procedures for PA intensity classification: score 84.65 %, versus up to 63.60 %. Errors tended to be brief and to appear around transients.

Conclusions: The application of ML techniques for pattern identification and temporal filtering allowed to merge accelerometry and HR data in a solid manner, and achieved markedly better recognition performances than the standard methods for PA intensity estimation.

 
  • References

  • 1 Fogelholm M. Physical activity, fitness and fatness: relations to mortality, morbidity and disease risk factors. A systematic review. Obes Rev 2010; 11 (Suppl. 03) 202-221. doi:10.1111/j.1467–789x.2009. 00653.x.
  • 2 Colberg SR, Sigal RJ, Fernhall B, Regensteiner JG, Blissmer BJ, Rubin RR. et al. Exercise and type 2 diabetes: The American College of Sports Medicine and the American Diabetes Association, Joint position statement. Diabetes Care 2010; 33 (Suppl. 12) e147-e167. doi:10.2337/dc10–9990.
  • 3 Thompson PD, Buchner D, Piña IL, Balady GJ, Williams MA, Marcus BH. et al. Exercise and physical activity in the prevention and treatment of atherosclerotic cardiovascular disease. Circulation 2003; 107 (Suppl. 24) 3109-3116. doi:10.1161/01.cir.0000075572.40158.77.
  • 4 Chavannes N, Vollenberg JJH, van Schayck CP, Wouters EFM. Effects of physical activity in mild to moderate COPD: A systematic review. Br J Gen Pract 2002; 52 (Suppl. 480) 574-578.
  • 5 Holmes MD. Physical activity and survival after breast cancer diagnosis. JAMA 2005; 293 (Suppl. 20) 2479-2486. doi:10.1001/jama.293.20.2479.
  • 6 García-García F, Kumareswaran K, Hovorka R, Hernando ME. Quantifying the acute changes in glucose with exercise in type 1 diabetes: A systematic review and meta-analysis. Sports Med 2015; 45 (Suppl. 04) 587-599. doi:10.1007/s40279–015–0302–2.
  • 7 Breton MD. Physical activity – The major unaccounted impediment to closed loop control. J Diabetes Sci Technol 2008; 2 (Suppl. 01) 169-174. doi:10.1177/193229680800200127.
  • 8 van Bon AC, Verbitskiy E, von Basum G, Hoekstra JBL, de Vries JH. Exercise in closed-loop control: A major hurdle. J Diabetes Sci Technol 2011; 5 (Suppl. 06) 1337-1341. doi:10.1177/193229681100500604.
  • 9 Valanou EM, Bamia C, Trichopoulou A. Methodology of physical activity and energy-expenditure assessment: A review. J Public Health 2006; 14 (Suppl. 02) 58-65. doi:10.1007/s10389–006–0021–0.
  • 10 Freedson PS, Miller K. Objective monitoring of physical activity using motion sensors and heart rate. Res Q Exerc Sport 2000; 71 (Suppl. 02) S21-S29.
  • 11 Gotshalk LA, Berger RA, Kraemer WJ. Cardiovascular responses to a high-volume continuous circuit resistance training protocol. J Strength Cond Res 2004; 18 (Suppl. 04) 760-764.
  • 12 Ainslie PN, Reilly T, Westerterp KR. Estimating human energy expenditure: A review of techniques with particular reference to doubly labelled water. Sports Med 2003; 33 (Suppl. 09) 683-698. doi:10.2165/00007256–200333090–00004.
  • 13 Plasqui G, Westerterp KR. Accelerometry and heart rate as a measure of physical fitness. Med Sci Sports Exerc 2006; 38 (Suppl. 08) 1510-1514. doi:10.1249/01.mss.0000228942.55152.84.
  • 14 Bassett DR, Rowlands A, Trost SG. Calibration and validation of wearable monitors. Med Sci Sports Exerc 2012; 44 Suppl S32-S38. doi:10.1249/mss.0b013e3182399cf7.
  • 15 Bächlin M, Plotnik M, Roggen D, Giladi N, Hausdorff JM, Tröster G. A wearable system to assist walking of Parkinson’s disease patients, Benefits and challenges of context-triggered acoustic cueing. Methods Inf Med 2010; 49 (Suppl. 01) 88-95. doi:10.3414/ME09–02–0003.
  • 16 Gietzelt M, Wolf KH, Kohlmann M, Marschollek M, Haux R. Measurement of accelerometry-based gait parameters in people with and without dementia in the field. Methods Inf Med 2013; 52 (Suppl. 04) 319-325. doi:10.3414/ME12–02–0009.
  • 17 Marschollek M, Rehwald A, Wolf KH, Gietzelt M. et al. Sensor-based fall risk assessment – An expert ‘to go’. Methods Inf Med 2011; 50 (Suppl. 05) 420-426. doi:10.3414/ME10–01–0040.
  • 18 Bao L, Intille SS. Activity recognition from user-annotated acceleration data. Proceedings of the 2nd International Conference PERVASIVE 2004: 1-17. doi:10.1007/978–3–540–24646–6_1.
  • 19 Dias A, Gorzelniak L, Schultz K, Wittmann M, Rudnik J. et al. Classification of exacerbation episodes in chronic obstructive pulmonary disease patients. Methods Inf Med 2014; 53 (Suppl. 02) 108-114. doi:10.3414/ME12–01–0108.
  • 20 Ermes M, Parkka J, Mantyjarvi J, Korhonen I. Detection of daily activities and sports with wearable sensors in controlled and uncontrolled conditions. IEEE Trans Inform Technol Biomed 2008; 12 (Suppl. 01) 20-26. doi:10.1109/titb.2007.899496.
  • 21 de Vries SI, Garre FG, Engbers LH, Hildebrandt VH, van Buuren S. Evaluation of neural networks to identify types of activity using accelerometers. Med Sci Sports Exerc 2011; 43 (Suppl. 01) 101-107. doi:10.1249/mss.0b013e3181e5797d.
  • 22 Lau HY, Tong KY, Zhu H. Support Vector Machine for classification of walking conditions using miniature kinematic sensors. Med Biol Eng Comput 2008; 46 (Suppl. 06) 563-573. doi:10.1007/s11517–008–0327-x.
  • 23 Gyllensten IC, Bonomi AG. Identifying types of physical activity with a single accelerometer: Evaluating laboratory-trained algorithms in daily life. IEEE Trans Biomed Eng 2011; 58 (Suppl. 09) 2656-2663. doi:10.1109/tbme.2011.2160723.
  • 24 Casale P, Pujol O, Radeva P. Human activity recognition from accelerometer data using a wearable device. Proceedings of the 5th Iberian Conference IbPRIA 2011: 289-296. doi:10.1007/978–3–642–21257–4_36.
  • 25 Lester J, Choudhury T, Borriello G. A practical approach to recognizing physical activities. Proceedings of the 4th International Conference PERVASIVE 2006: 1-16. doi:10.1007/11748625_1.
  • 26 Tapia EM, Intille SS, Haskell W, Larson K, Wright J, King A. et al. Real-time recognition of physical activities and their intensities using wireless accelerometers and a heart rate monitor. Proceedings of the 11th IEEE International Symposium on Wearable Computers 2007 doi:10.1109/iswc.2007. 4373774.
  • 27 Li M, Rozgic V, Thatte G, Lee S, Emken A, Anna-varam M. et al. Multimodal physical activity recognition by fusing temporal and cepstral information. IEEE Trans Neural Syst Rehab Eng 2010; 18 (Suppl. 04) 369-380. doi:10.1109/tnsre.2010.2053217.
  • 28 Reiss A, Stricker D. Aerobic activity monitoring: towards a long-term approach. Univ Access Inf Soc 2013; 13 (Suppl. 01) 101-114. doi:10.1007/s10209–013–0292–5.
  • 29 Albinali F, Intille S, Haskell W, Rosenberger M. Using wearable activity type detection to improve physical activity energy expenditure estimation.. Proceedings of the 12th International Conference on Ubiquitous Computing 2010. doi:10.1145/ 1864349.1864396.
  • 30 Atallah L, Leong JJH, Lo B, Yang GZ. Energy expenditure prediction using a miniaturized ear-worn sensor. Med Sci Sports Exerc 2011; 43 (Suppl. 07) 1369-1377. doi:10.1249/mss.0b013e3182093014.
  • 31 Liu S, Gao RX, John D, Staudenmayer J, Freedson PS. SVM-based multi-sensor fusion for free-living physical activity assessment. Proceedings of the 33rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society EMBS. 2011: 3188-3191. doi:10.1109/iembs.2011. 6090868.
  • 32 Altini M, Penders J, Vullers R, Amft O. Estimating energy expenditure using body-worn accelerometers: A comparison of methods, sensors number and positioning. IEEE J Biomed Health Inform 2015; 19 (Suppl. 01) 219-226. doi:10.1109/jbhi. 2014.2313039.
  • 33 Lin CW, Yang YTC, Wang JS, Yang YC. A wearable sensor module with a neural-network-based activity classification algorithm for daily energy expenditure estimation. IEEE Trans Inform Technol Biomed 2012; 16 (Suppl. 05) 991-998. doi:10.1109/titb. 2012.2206602.
  • 34 Freedson PS, Lyden K, Kozey-Keadle S, Stauden-mayer J.. Evaluation of artificial neural network algorithms for predicting METs and activity type from accelerometer data: validation on an independent sample. J Appl Physiol 2011; 111 (Suppl. 06) 1804-1812. doi:10.1152/japplphysiol.00309.2011.
  • 35 Trost SG, Wong WK, Pfeiffer KA, Zheng Y. Artificial neural networks to predict activity type and energy expenditure in youth. Med Sci Sports Exerc 2012; 44 (Suppl. 09) 1801-1809. doi:10.1249/mss. 0b013e318258ac11.
  • 36 Bassett DR, Ainsworth BE, Swartz AM, Strath SJ, O’Brien WL, King GA. Validity of four motion sensors in measuring moderate intensity physical activity. Med Sci Sports Exerc 2000; 32 (Suppl) S471-480. doi:10.1097/00005768–200009001–00006.
  • 37 Welk GJ, Schaben JA, Morrow JRJ. Reliability of accelerometry-based activity monitors: a generalizability study. Med Sci Sports Exerc 2004; 36 (Suppl. 09) 1637-1645.
  • 38 Sasaki JE, John D, Freedson PS. Validation and comparison of ActiGraph activity monitors. J Sci Med Sport 2011; 14 (Suppl. 05) 411-416. doi:10.1016/j.jsams.2011.04.003.
  • 39 Le Masurier GC, Tudor-Locke C. Comparison of pedometer and accelerometer accuracy under controlled conditions. Med Sci Sports Exerc 2003; 35 (Suppl. 05) 867-871. doi:10.1249/01.mss.0000064996.63632.10.
  • 40 Chen KY, Bassett DR. The technology of accelerometry-based activity monitors: Current and future. Med Sci Sports Exerc 2005; 37 (Suppl) S490-S500. doi:10.1249/01.mss.0000185571.49104.82.
  • 41 ActiGraph.. ActiLife 5 Analysis Software Suite for ActiGraph devices (User manual–Revision H). 2010
  • 42 Zapico AG, Benito PJ, González-Gross M, Peinado AB, Morencos E, Romero B. et al. Nutrition and physical activity programs for obesity treatment (PRONAF study): Methodological approach of the project. BMC Public Health 2012; 12 (Suppl. 01) 1100. doi:10.1186/1471–2458–12–1100.
  • 43 Byrne NM, Hill AP, Hunter GR, Weinsier RL, Schutz Y. Metabolic equivalent: one size does not fit all. J Appl Physiol 2005; 99 (Suppl. 03) 1112-1119.
  • 44 Wasserman K, Hansen JE, Sue DY, Whipp BJ, Casaburi R. Principles of exercise testing and interpretation. 2nd ed. Lippincott Williams & Wilkins; 1994
  • 45 Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C. et al. Physical activity and public health. A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 1995; 273 (Suppl. 05) 402-407.
  • 46 Ainsworth BE, Haskell WL, Whitt MC, Irwin ML, Swartz AM, Strath SJ. et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc 2000; 32 (Suppl) S498-S516. doi:10.1097/00005768–200009001–00009.
  • 47 Benito PJ, Bermejo LM, Peinado AB, López-Plaza B, Cupeiro R, Szendrei B. et al. Change in weight and body composition in obese subjects following a hypocaloric diet plus different training programs or physical activity recommendations. J Appl Physiol 2015; 118 (Suppl. 08) 1006-1013. doi:10.1152/japplphysiol.00928.2014.
  • 48 Preece SJ, Goulermas JY, Kenney LPJ, Howard D, Meijer K, Crompton R. Activity identification using body-mounted sensors – A review of classification techniques. Physiol Meas 2009; 30 (Suppl. 04) R1-R33. doi:10.1088/0967–3334/30/4/r01.
  • 49 Duda RO, Hart PE, Stork DG. Pattern Classification. Wiley; 2000
  • 50 Zumel N, Mount J. Practical Data Science. with R. Manning Publications; 2014
  • 51 Freedson PS, Melanson E, Sirard J. Calibration of the Computer Science and Applications, Inc. accelerometer. Med Sci Sports Exerc. 1998 30. 05 777-781. doi:10.1097/00005768–199805000–00021.
  • 52 Rencher AC, Christensen WF. Methods of Multivariate Analysis. John Wiley & Sons; 2012
  • 53 Kohavi R, John GH. Wrappers for feature subset selection. Artificial Intelligence 1997; 97 (Suppl. 01) 273-324. doi:10.1016/s0004–3702(97)00043-x.
  • 54 Guyon I, Elisseeff A. An introduction to variable and feature selection. J Mach Learn Res 2003; 3: 1157-1182.
  • 55 Peng H, Long F, Ding C. Feature selection based on mutual information criteria of max-dependency, max-relevance, and min-redundancy. IEEE Trans Pattern Anal Machine Intell 2005; 27 (Suppl. 08) 1226-1238. doi:10.1109/tpami.2005.159.
  • 56 Ding C, Peng H. Minimum redundancy feature selection from microarray gene expression data. J Bioinform Comput Biol 2005; 03 (Suppl. 02) 185-205. doi:10.1142/s0219720005001004.
  • 57 Pober DM, Staudenmayer J, Raphael C, Freedson PS. Development of novel techniques to classify physical activity mode using accelerometers. Med Sci Sports Exerc 2006; 38 (Suppl. 09) 1626-1634. doi:10.1249/01.mss.0000227542.43669.45.
  • 58 He J, Li H, Tan J. Real-time daily activity classification with wireless sensor networks using Hidden Markov Model. Proceedings of the 29th Annual International Conference of the IEEE Engineering in Medicine and Biology Society EMBS 2007 doi:10.1109/iembs.2007.4353008.
  • 59 Mannini A, Sabatini A. M.. Accelerometry-based classification of human activities using Markov modeling. Comput Intell Neurosci 2011; 2011: 1-10. doi:10.1155/2011/647858.
  • 60 Rabiner L. A tutorial on Hidden Markov Models and selected applications in speech recognition. Proceedings of the IEEE 1989; 77 (Suppl. 02) 257-286. doi:10.1109/5.18626.
  • 61 Forman G, Scholz M. Apples-to-apples in cross-validation studies. ACM SIGKDD Explor Newsl 2010; 12 (Suppl. 01) 49. doi:10.1145/1882471.1882479.
  • 62 Crouter SE, Churilla JR, Bassett DR. Estimating energy expenditure using accelerometers. Eur J Appl Physiol 2006; 98 (Suppl. 06) 601-612. doi:10.1007/ s00421–006–0307–5.
  • 63 Brage S, Brage N, Franks PW, Ekelund U, Wareham NJ.. Reliability and validity of the combined heart rate and movement sensor Actiheart. Eur J Clin Nutr 2005; 59 (Suppl. 04) 561-570. doi:10.1038/ sj.ejcn.1602118.
  • 64 American College of Sports Medicine.. ACSM’s guidelines for exercise testing and prescription. Lippincott Williams & Wilkins; 2013
  • 65 Wareham NJ, Rennie KL. The assessment of physical activity in individuals and populations: why try to be more precise about how physical activity is assessed?. Int J Obes Relat Metab Disord 1998; 22 (Suppl. 02) S30-38.
  • 66 Altini M, Penders J, Vullers R, Amft O. Automatic heart rate normalization for accurate energy expenditure estimation. Methods Inf Med 2014; 53 (Suppl. 05) 382-388. doi:10.3414/ME13–02–0031.
  • 67 Froelicher VF, Myers J. Exercise and the heart. 5th ed. Saunders; 2006
  • 68 Coote JH. Recovery of heart rate following intense dynamic exercise. Exp Physiol 2010; 95 (Suppl. 03) 431-440. doi: 10.1113/expphysiol.2009.047548.
  • 69 Hall KD, Heymsfield SB, Kemnitz JW, Klein S, Schoeller DA, Speakman JR. Energy balance and its components: implications for body weight regulation. Am J Clin Nutr 2012; 95 (Suppl. 04) 989-994. doi:10.3945/ajcn.112.036350.