1st International Workshop on Lifelogging for Pervasive Health

Research Article

A Sensor Device for Automatic Food Lifelogging that is Embedded in Home Ceiling Light: A Preliminary Investigation

Download831 downloads
  • @INPROCEEDINGS{10.4108/icst.pervasivehealth.2013.252128,
        author={Takuya Maekawa},
        title={A Sensor Device for Automatic Food Lifelogging that is Embedded in Home Ceiling Light: A Preliminary Investigation},
        proceedings={1st International Workshop on Lifelogging for Pervasive Health},
        publisher={IEEE},
        proceedings_a={LIFELOGGING},
        year={2013},
        month={5},
        keywords={food lifelog sensor},
        doi={10.4108/icst.pervasivehealth.2013.252128}
    }
    
  • Takuya Maekawa
    Year: 2013
    A Sensor Device for Automatic Food Lifelogging that is Embedded in Home Ceiling Light: A Preliminary Investigation
    LIFELOGGING
    IEEE
    DOI: 10.4108/icst.pervasivehealth.2013.252128
Takuya Maekawa,*
    *Contact email: maekawa@ist.osaka-u.ac.jp

    Abstract

    Due to the recent proliferation of digital cameras and smart phones with camera devices, many researchers have attempted to store and analyze photographs (images) that capture a user's meal. Simply storing photographs of meals before eating can encourage weight loss. Also, by analyzing the images, some researchers attempt to estimate the nutritional composition of the meal. However, these approaches rely on images manually photographed by the user. So, when the user forgets to take a picture of his/her meal, the information related to the meal will be lost. In this paper, we propose and design a sensor device with a camera that automatically takes a photograph of a user's meal. The device is attached to a ceiling light in a dining room of the user's house. So, the device is supplied with electricity from the light. Also, the device has a camera and uses it to capture the dining table under the ceiling light. With this device, we can automatically and continually take photographs during the user's mealtime. Here the problem is how to determine a representative photograph of the user's meal from the continually captured images. In this paper, we investigate how to find the representative photograph captured during the mealtime by using the complexity of an image.