Facio is a visualisation project that was birthed from aspects of our Honours Project which is based on face perception and recognition. We are taking human faces and, by means of facial metrics, visualising them in two simple and meaningful graphical representations. The user simply supplies one photograph of their face, and possibly a second one of another face if they want a direct comparison. They then place a number of predefined markers on strategic points that map onto the uploaded photo. It is from these markers that the system gets input data. The measurements are either converted into ratios or normalised to compensate for pictures with different depth values. The output will be two separate visualisations:

Symmetry: this will display the normalised lengths of features from the centre of the face. This is for comparing face symmetry.

Feature comparison: this will compare two different faces using a heptagonal shape (representing the seven facial metrics that we selected). This visualisation will highlight the differences in ratios between the two faces.

These visualisations could be used by different users. They could just be random users who want to compare their face to someone elses, or to test their facial symmetry. These would be our casual users who are just using the system for fun. If we were to extend the system to store all of the input data into a database, our system can be used by researchers who have interests in facial perception. If lots of data was captured from different individuals, a good average output visualisation can be produced. This can be used, for example, to map average features of people in certain groups such as gender, race, nationality, etc.

The work for this project was divided up into three sections. Everyone was involved when it came to brainstorming initial designs. Once initial designs were created, the programming of the system was divided into:

GUI input and integration of visualizations - Dane Mackier

Symmetry Visualization - Grant van Helsdingen

Facial Ratios Visualization - Laurent Daniels

Once the programming was done, Dane built the website hosted on his public HTML folder and every group member contributed to writing the report, adding sections pertinent to the work they did.