Advertisement

Building Parsimonious SVM Models for Chewing Detection and Adapting Them to the User

  • Iason Karakostas
  • Vasileios PapapanagiotouEmail author
  • Anastasios Delopoulos
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10590)

Abstract

Monitoring of eating activity is a well-established yet challenging problem. Various sensors have been proposed in the literature, including in-ear microphones, strain sensors, and photoplethysmography. Most of these approaches use detection algorithms that include machine learning; however, a universal, non user-specific model is usually trained from an available dataset for the final system. In this paper, we present a chewing detection system that can adapt to each user independently using active learning (AL) with minimal intrusiveness. The system captures audio from a commercial bone-conduction microphone connected to an Android smart-phone. We employ a state-of-the-art feature extraction algorithm and extend the Support Vector Machine (SVM) classification stage using AL. The effectiveness of the adaptable classification model can quickly converge to that achieved when using the entire available training set. We further use AL to create SVM models with a small number of support vectors, thus reducing the computational requirements, without significantly sacrificing effectiveness. To support our arguments, we have recorded a dataset from eight participants, each performing once or twice a standard protocol that includes consuming various types of food, as well as non-eating activities such as silent and noisy environments and conversation. Results show accuracy of 0.85 and F1 score of 0.83 in the best case for the user-specific models.

Keywords

Active learning Dietary monitoring Chewing detection Wearable sensors 

Notes

Acknowledgements

The work leading to these results has received funding from the European Communitys Health, demographic change and well-being Programme under Grant Agreement No. 727688, 01/12/2016 - 30/11/2020.

References

  1. 1.
    Amft, O., Stäger, M., Lukowicz, P., Tröster, G.: Analysis of chewing sounds for dietary monitoring. In: Beigl, M., Intille, S., Rekimoto, J., Tokuda, H. (eds.) UbiComp 2005. LNCS, vol. 3660, pp. 56–72. Springer, Heidelberg (2005).  https://doi.org/10.1007/11551201_4 CrossRefGoogle Scholar
  2. 2.
    Archer, E., Hand, G.A., Blair, S.N.: Validity of us nutritional surveillance: national health and nutrition examination survey caloric energy intake data, 1971–2010. PloS ONE 8(10), e76632 (2013)CrossRefGoogle Scholar
  3. 3.
    Fontana, J.M., Farooq, M., Sazonov, E.: Automatic ingestion monitor: a novel wearable device for monitoring of ingestive behavior. IEEE Trans. Biomed. Eng. 61(6), 1772–1779 (2014)CrossRefGoogle Scholar
  4. 4.
    Kyritsis, K., Tatli, C.L., Diou, C., Delopoulos, A.: Automated analysis of in meal eating behavior using a commercial wristband IMU sensor. In: 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (July 2017, to appear online)Google Scholar
  5. 5.
    Papapanagiotou, V., Diou, C., Delopoulos, A.: Chewing detection from an in-ear microphone using convolutional neural networks. In: 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (July 2017, to appear online)Google Scholar
  6. 6.
    Papapanagiotou, V., Diou, C., Lingchuan, Z., van den Boer, J., Mars, M., Delopoulos, A.: Fractal nature of chewing sounds. In: Murino, V., Puppo, E., Sona, D., Cristani, M., Sansone, C. (eds.) ICIAP 2015. LNCS, vol. 9281, pp. 401–408. Springer, Cham (2015).  https://doi.org/10.1007/978-3-319-23222-5_49 CrossRefGoogle Scholar
  7. 7.
    Papapanagiotou, V., Diou, C., Zhou, L., van den Boer, J., Mars, M., Delopoulos, A.: A novel chewing detection system based on PPG, audio and accelerometry. IEEE J. Biomed. Health Inf. 21, 607–618 (2016)CrossRefGoogle Scholar
  8. 8.
    Sazonov, E., Schuckers, S., Lopez-Meyer, P., Makeyev, O., Sazonova, N., Melanson, E.L., Neuman, M.: Non-invasive monitoring of chewing and swallowing for objective quantification of ingestive behavior. Physiol. Meas. 29(5), 525 (2008)CrossRefGoogle Scholar
  9. 9.
    Sazonov, E.S., Fontana, J.M.: A sensor system for automatic detection of food intake through non-invasive monitoring of chewing. IEEE Sens. J. 12(5), 1340–1348 (2012)CrossRefGoogle Scholar
  10. 10.
    Shuzo, M., Komori, S., Takashima, T., Lopez, G., Tatsuta, S., Yanagimoto, S., Warisawa, S., Delaunay, J.J., Yamada, I.: Wearable eating habit sensing system using internal body sound. J. Adv. Mech. Des. Syst. Manuf. 4(1), 158–166 (2010)CrossRefGoogle Scholar
  11. 11.
    Tong, S., Koller, D.: Support vector machine active learning with applications to text classification. J. Mach. Learn. Res. 2, 45–66 (2001)zbMATHGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Iason Karakostas
    • 1
  • Vasileios Papapanagiotou
    • 1
    Email author
  • Anastasios Delopoulos
    • 1
  1. 1.Multimedia Understanding Group, Department of Electrical and Computer EngineeringAristotle University of ThessalonikiThessalonikiGreece

Personalised recommendations