Skip to main content
SpringerLink
Log in

Unsupervised template discovery in activity recognition using the Gamma Growing Neural Gas algorithm

  • Focus
  • Published:
Soft Computing Aims and scope Submit manuscript

Abstract

Activity recognition is gaining a lot of interest given its direct use in applications like ambient assisted living and has been empowered by the increasing ubiquity of sensors (e.g., clothes, smartphones, watches). The machine learning approach to activity recognition consists on finding the signatures characterizing the activities to be recognized, with the hope of identifying them (pattern matching) within the stream of sensor data. The finding of those signatures can be very complex, thus many approaches deal with the streams of sensor data by segmenting them into sections or “time-windows”, before processing them by a feature extraction procedure. The problem then concerns the association of features to class labels. In this paper, we propose the use of the Gamma Growing Neural Gas algorithm to unsupervisely discover templates in a recording containing gestures performed by a person in a home environment. The system is able to do vector quantization from the time-series of data coming from one accelerometer, and finds salient patterns (e.g., templates) in the signal. These templates integrate information not only from single time-windows but do consider the recent history of the incoming signal (e.g., multiple time-windows). Those templates are then associated to activity classes by supervised learning. Our experiments show that the resulting performance is better than previous benchmarks of the same database.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3

Similar content being viewed by others

Notes

  1. A collection of snapshots of the person or images of an avatar performing the activity.

  2. Since a longer portion of the signal is available, no features characterizing the dynamics of the signal within long time-windows are needed.

  3. The IMU includes a three-axial accelerometer and a three-axial gyroscope sampled at 102.4 Hz.

  4. The dataset contains the execution of other strokes e.g., forehand and backhand topspin, forehand and backhand block.

  5. See (Estévez and Hernández 2011) for the description of each parameter.

  6. We computed the best matching unit (BMU) for each point in the dataset.

  7. The performance of the detection of the null class was not evaluated since this class is much more frequent in the dataset. The performance of the system is overestimated if the null class is considered in the evaluation of the weighted \(F1\)-score.

References

  • Aggarwal JK, Ryoo MS (2011) Human activity analysis: a review. ACM Comput Surv 43(3):1–16. doi:10.1145/1922649.1922653 (ISSN 0360–0300)

    Article  Google Scholar 

  • Amft O, Tröster G (2008) Recognition of dietary activity events using on-body sensors. Artif Intell Med 42(2):121–136. doi:10.1016/j.artmed.2007.11.007 (ISSN 0933–3657)

    Article  Google Scholar 

  • Baños O, Damas M, Pomares H, Rojas I, Tóth MA, Amft O (2012) A benchmark dataset to evaluate sensor displacement in activity recognition. In: Proceedings of the 2012 ACM Conference on Ubiquitous Computing, UbiComp ’12, New York, ACM, pp. 1026–1035. doi:10.1145/2370216.2370437 (ISBN 978-1-4503-1224-0)

  • Baños O, Damas M, Pomares H, Rojas I (2013) Activity recognition based on a multi-sensor meta-classifier. In: Rojas I, Joya G, Cabestany J (eds) Advances in computational Intelligence, volume 7903 of Lecture Notes in Computer Science. Springer, Heidelberg, pp. 208–215. doi:10.1007/978-3-642-38682-4_24 (ISBN 978-3-642-38681-7)

  • Bache K, Lichman M (2013) UCI machine learning repository. http://archive.ics.uci.edu/ml

  • Browne G, Berry E, Narinder K, Hodges S, Smyth G, Watson P, Wood K (2011) Sensecam improves memory for recent events and quality of life in a patient with memory retrieval difficulties. Memory 19(7):713–722

    Article  Google Scholar 

  • Bulling A, Blanke U, Schiele B (2014) A tutorial on human activity recognition using body-worn inertial sensors. ACM Comput Surv 46:1–33

  • Chavarriaga R, Bayati H, Del Millán J (2013) Unsupervised adaptation for acceleration-based activity recognition: Robustness to sensor displacement and rotation. Pers Ubiquitous Comput 17(3):479–490. doi:10.1007/s00779-011-0493-y (ISSN 1617–4909)

  • Chen L, Hoey J, Nugent CD, Cook DJ, Yu Z (2012) Sensor-based activity recognition. Syst Man Cybern Part C Appl Rev IEEE Trans 42(6):790–808. doi:10.1109/TSMCC.2012.2198883 (ISSN 1094–6977)

    Article  Google Scholar 

  • de Vries B, Principe JC (1992) The gamma model, a new neural model for temporal processing. Neural Netw 5(4):565–576. doi:10.1016/S0893-6080(05)80035-8 (ISSN 0893–6080)

    Article  Google Scholar 

  • Delachaux B, Rebetez J, Perez-Uribe A, Satizábal HF (2013) Indoor activity recognition by combining one-vs-all neural network classifiers exploiting wearable and depth sensors. In: Proceedings of the International Work-Conference. on Artificial Neural Networks. Springer, Heidelberg, pp 216–223

  • Estévez PA, Hernández R (2011) Gamma-filter self-organizing neural networks for time series analysis. In: Laaksonen J, Honkela T (eds) Advances in self-organizing maps, volume 6731 of Lecture Notes in Computer Science. Springer, Berlin Heidelberg, pp 151–159. doi:10.1007/978-3-642-21566-7_15 (ISBN 978-3-642-21565-0)

  • Fritzke B (1995) A growing neural gas network learns topologies. In: Tesauro G, Touretzky DS, Leen TK (eds) Advances in Neural Information Processing Systems 7. MIT Press, Cambridge, pp 625–632

    Google Scholar 

  • Hartmann B, Link N (2010) Gesture recognition with inertial sensors and optimized dtw prototypes. In: Systems Man and Cybernetics (SMC), 2010 IEEE International Conferenc. pp 2102–2109. doi:10.1109/ICSMC.2010.5641703

  • Hondori HM, Khademi M, Lopes CV (2012) Monitoring intake gestures using sensor fusion (microsoft kinect and inertial sensors) for smart home tele-rehab setting. In: IEEE HIC 2012 Engineering in Medicine and Biology Society Conference on Healthcare Innovation.

  • Kasteren TLM, Englebienne G, Kröse BJA (2011) Human activity recognition from wireless sensor network data: Benchmark and software. In: Chen L, Nugent CD, Biswas J, Hoey J (eds) Activity recognition in pervasive intelligent environments, volume 4 of Atlantis Ambient and Pervasive Intelligence. Atlantis Press, pp. 165–186. doi:10.2991/978-94-91216-05-3_8 (ISBN 978-90-78677-42-0)

  • Kepski M, Kwolek B (2012) Fall detection on embedded platform using kinect and wireless accelerometer. In: Proceedings of the 13th International Conference on Computers Helping People with Special Needs— Volume Part II, ICCHP’12. Springer, Heidelberg, pp. 407–414. doi:10.1007/978-3-642-31534-3_60 (ISBN 978-3-642-31533-6)

  • Leppanen J, Eronen A (2013) Accelerometer-based activity recognition on a mobile phone using cepstral features and quantized gmms. In: Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference. pp 3487–3491. doi:10.1109/ICASSP.2013.6638306

  • Lukowicz P, Pirkl G, Bannach D, Wagner F, Calatroni A, Förster K, Holleczek T, Rossi M, Roggen D, Tröster G, Doppler J, Holzmann C, Riener A, Ferscha A, Chavarriaga R (2010) Recording a complex, multi modal activity data set for context recognition. In: ARCS Workshops, pp 161–166

  • Nguyen-Dinh L-V , Roggen D, Calatroni A, Troster G (2012) Improving online gesture recognition with template matching methods in accelerometer data. In: Intelligent Systems Design and Applications (ISDA), 2012 12th International Conference on, pp 831–836. doi:10.1109/ISDA.2012.6416645

  • Ni B, Wang G, Moulin P (2013) Rgbd-hudaact: A color-depth video database for human daily activity recognition. In: Fossati A, Gall J, Grabner H, Ren X, Konolige K (eds) Consumer depth cameras for computer vision, Advances in computer vision and pattern recognition. Springer, London, pp 193–208. doi:10.1007/978-1-4471-4640-7_10 (ISBN 978-1-4471-4639-1)

  • Lara ÓD, Pérez AJ, Labrador MA, Posada JD (2012) Centinela: a human activity recognition system based on acceleration and vital sign data. Pervasive Mobile Comput 8(5):717–729. doi:10.1016/j.pmcj.2011.06.004 (ISSN 1574–1192)

    Article  Google Scholar 

  • Principe JC, de Vries B, de Oliveira PG (1993) The gamma-filter-a new class of adaptive iir filters with restricted feedback. Trans Sig Proc 41(2):649–656. doi:10.1109/78.193206 (ISSN 1053–587X)

  • Rawassizadeh R, Tomitsch M, Wac K, Min Tjoa A (2013) Ubiqlog: a generic mobile phone-based life-log framework. Personal and Ubiquitous Computing, 17(4):621–637. doi:10.1007/s00779-012-0511-8 (ISSN 1617–4909)

  • Rebetez J, Satizábal HF, Perez-Uribe A (2013) Reducing user intervention in incremental activityrecognition for assistive technologies. In: Proceedings of the 2013 International Symposium on Wearable Computers, ISWC ’13. New York. ACM. pp 29–32. doi:10.1145/2493988.2494350 (ISBN 978-1-4503-2127-3)

  • Van Rijsbergen CJ (1979) Information Retrieval. 2nd edn. Butterworth-Heinemann, Newton (ISBN 0408709294)

  • Roggen D, Calatroni A, Rossi M, Holleczek T, Förster K, Tröster G, Lukowicz P, Bannach D, Pirkl G, Ferscha A, Doppler J, Holzmann C, Kurz M, Holl G, Chavarriaga R, Sagha H, Bayati H, Creatura M, Millán JR (2010) Collecting complex activity datasets in highly rich networked sensor environments. In: Networked Sensing Systems (INSS), 2010 Seventh International Conference on, pp 233–240. doi:10.1109/INSS.2010.5573462

  • Sagha H, Digumarti ST, Millán JR, Chavarriaga R, Calatroni A, Roggen D, Tröster G (2011) Benchmarking classification techniques using the Opportunity human activity dataset. In: 2011 Ieee International Conference On Systems, Man, And Cybernetics (Smc), IEEE International Conference on Systems Man and Cybernetics Conference Proceedings. IEEE service center, Piscataway, pp 36–40

  • Satizábal HF, Rebetez J , Perez-Uribe A (2013) Semi-supervised discovery of time-series templates for gesture spotting in activity recognition. In: Proceedings of the 2nd International Conference in Pattern Recognition Applications and Methods. pp 573–576

  • Shotton J, Sharp T, Kipman A, Fitzgibbon A, Finocchio M, Blake A, Cook M, Moore R (2013) Real-time human pose recognition in parts from single depth images. Commun ACM 56(1):116–124. doi:10.1145/2398356.2398381 (ISSN 0001–0782)

  • Stiefmeier T, Roggen D, Ogris G, Lukowicz P, Lukowicz P (2008) Wearable activity tracking in car manufacturing. Pervasive Comput IEEE 7(2):42–50. doi:10.1109/MPRV.2008.40 (ISSN 1536–1268)

  • van Kasteren TLM, Englebienne G, Kröse BJA (2011) Hierarchical activity recognition using automatically clustered actions. In: Proceedings of the Second international conference on Ambient Intelligence, Am I’11. Springer, Heidelberg, pp 82–91. doi:10.1007/978-3-642-25167-2_9 (ISBN 978-3-642-25166-5)

  • Xue Y, Jin L (2010) A naturalistic 3d acceleration-based activity dataset amp; benchmark evaluations. In: Systems Man and Cybernetics (SMC), 2010 IEEE International Conference. pp 4081–4085. doi:10.1109/ICSMC.2010.5641790

Download references

Acknowledgments

This work was funded by the Hasler Foundation. We thank our colleagues from the Wearable Computing Lab. ETH, for useful and interesting discussions.

Author information

Authors and Affiliations

  1. IICT, HEIG-VD, University of Applied Sciences Western Switzerland (HES-SO), Delémont, Switzerland

    Héctor F. Satizábal & Andres Perez-Uribe

Authors
  1. Héctor F. Satizábal
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Andres Perez-Uribe
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Héctor F. Satizábal.

Additional information

Communicated by I. R. Ruiz.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Satizábal, H.F., Perez-Uribe, A. Unsupervised template discovery in activity recognition using the Gamma Growing Neural Gas algorithm. Soft Comput 19, 2435–2445 (2015). https://doi.org/10.1007/s00500-014-1499-y

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00500-014-1499-y

Keywords

Search

Navigation