Speaker:
Maria Antonietta Pascali
- Signals & Images lab, ISTI CNR Pisa
Tuesday, July 7, 2015
at
5:00 PM
16:45 rinfresco; 17:00 inizio seminario
Our objective is to infer weight variations of a subject from 3D facial scans acquired over time, to support self-monitoring and wellbeing improvement.
Starting from the state of the art, we defined a set of features: simple linear and planar measurements, sectional features, and more complex diagrams characterizing the homological structure of a facial mesh. We compared these features on a synthetic dataset of 3D faces with respect to the capability of encoding the subject's weight variation, and the easiness of implementation in the real setting (such as the quality of scans, and the facial landmarking dependance).
This work is carried out in the framework of the EU project SEMEOTICONS, devoted to self monitoring for wellbeing improvement by extracting facial signs of cardio-metabolic risk from video sequences and 3D scans.
