Thanks to the Erasmus Mundus I had the opportunity to work in a foreign lab, learning new stuffs, improving my knowledge, seeing things from a different perspective.
In this period, I worked on micro-behavior and human action recognition in the context of domestic house-user interaction. This is one of the main topic of this lab, which developed a SMART-HOME, a home in which various sensing technologies are integrated and used to apply various techniques of machine learning aimed at improving the quality of life people, especially elder ones. The idea is that the same house would be able to help the user during the day, suggesting action, preventing the waste of energy, monitoring the state of the user.
My research started with the deployment in this environment of the 3D distributed monitoring system derived from the AUSILIA project. From this system we are able to get the 3D skeleton of the used in key areas of the SMART-HOME, information useful for the identification of so called “micro/atomic” actions. An example is the set of actions related to the cooking: cutting, mixing, heating etc. If recognized, these action would allow the identification of the macro action of cocking, but also the possible identification of the recipe performed and its correctness in the steps.
The system is composed by 5 NUCs settled in various positions in the kitchen and living room of he SMART-HOME. We collected 12 days of subjects. We have now more than 100Gb of data.
Now it is time to proceed in the machine learning phase!
Not only research.. also sushi!
The I found amazing new friends and colleagues, first of all Yasumoto-Sensei, Arakawa-Sensei and Mizumoto-Sensei, so as all the students! Thanks you for all! See you soon!