Inhalt des Dokuments
Finding Critical Features for Difficulty Adaptation in Tablet-based Serious Games for Dementia
LOCATION: TEL, Auditorium 3 (20th floor), Ernst-Reuter-Platz 7, 10587 Berlin
Date/Time: 30.01.2017, 14:00-14:30
SPEAKER: Jeehoon Cha (TU Berlin)
Currently, there is no medication for dementia which stops or slows down their underlying neuro-degeneration process directly. Instead, non-pharmacologic treatments such as exercising, gardening, and playing games can be effective for improving quality of life of people with dementia. Some digital tools are also studied for the use of people with dementia in this context. PflegeTab is a tablet-based application organized for the use of people with dementia in this context. Its main purposes include improving quality-of-life of dementia people and having positive impacts on their mental states. To this end, it contains four kinds of applications: cognitive games, everyday activity training games, emotional applications, and a picture album as a biographical application. In this thesis, we find critical features for difficulty adaptation in the games of PflegeTab. Afterwards, we present prototypical difficulty adaptation mechanism and show that it can be performed seamlessly. Additionally, we also track the feasibility of mood classification of the users in the games with the same approach used in difficulty level classification. As a result, we identify that 1) touch speed features are critical for mood classification, and 2) time spent for a task reflects difficulty of the task. Especially, for mood classification, we could achieve 78.4% of accuracy.