I’ve been thinking further about the possibilities of Tim’s wall of faces as a finding aid, as something to help both locate archival documents and to understand their context.

The series we used in our test (ST84/1) was one in which we knew there was a very high percentage of photographs. Each item contains ten certificates, most of which have both a front and profile portrait attached. There is a small amount of other paperwork included in some files, but not a whole lot. We therefore knew what sorts of things we were going to get back.

But what about if we apply the same facial detection technology to a series in which we aren’t so sure of the photographic content? Unfortunately, Tim’s current laptop isn’t up to the task of doing all the grunt work (donations, anyone?), but here’s what I reckon might happen when we are able to move on to other series.

With series like SP42/1 and B13, which hold applications for CEDTs and similar records, I know that there are photographs in many, even most, of the personal case files. (B13 is complicated because it also contains other Customs files that don’t relate to individuals and don’t relate to the administration of the Immigration Restriction Act.) Because files might hold applications for a family, or a parent and child/ren, or an uncle and nephew, or siblings, you don’t always know from the item title exactly who the file relates to. Also, those who were Australian born did not necessarily apply for CEDTs since they could travel using their birth certificates as proof of their right to return, meaning that they don’t appear in CEDT series like ST84/1.

It was usual practice, though, to supply photographs of each person who was travelling (whether on a CEDT or not), and so by extracting those photographs, you would be able to have a better impression about who files related to. Of course, for files that are digitised (or even not) you could go through each one individually (which I’ve done, believe me…), but think how much more fun it would be to scroll through a wall of beautiful faces!

With B13 it would also be useful because there is no separate series of CEDTs; they are mixed in with the application/case files. Facial detection could be a way of extracting the forms themselves from the larger files.

My main research interest is in families, and women, and children – and we know that women are often hidden in archives because of bureaucratic systems which gave priority to the men in their lives. Although there are many White Australia records which relate to individual women and children, they can be lost in files organised and catalogued under the names of husbands and fathers. But scroll through a wall of mostly male faces, and the women and children just leap out at you!

I’m feeling a bit impatient, really, about running SP42/1 and B13 through Tim’s facial detection script. There are so many, so very interesting possibilities.