Computers and Curators: The Intersection in Image Curation

Is it curator against the machine? Or curator and the machine in harmony? Illustration by Shauna Colgan

There is no denying that computer technology has revolutionised how we recreate and analyse artefacts. There are now an array of technologies available allowing the reproduction of artefacts in non-invasive ways. Similarly, computer technology has found a home in image preservation with the digitisation of photographs. However, digitised photographs still require curation and this is where computational imaging analysis steps in. Such technology has the ability to recognise common features in photographs – ranging from relatively simple things like colours to more complex objects such as faces and landscape features – and group them together accordingly. While some may feel that this removes the human element from image analysis this would be a misinformed viewpoint. Computational imaging analysis still requires a human curator, and the job of the human curator is made easier thanks to computational imaging analysis. Hence, when building image databases computational analysis and human curators must work together for optimal results.

There are an undoubted number of benefits to using computational imaging analysis. Software has the ability to parse through a large dataset in minutes and categorise different images according to the features present in the image and the metadata information encoded. This is a job which could take a human curator days or potentially weeks depending on the size of the dataset and how much information in given. Imaging analysis software also has the ability to make developing databases searchable and potentially “live” much faster, even while the curator is still conducting contextual research into the images and the subjects present. These capabilities ensure that computational imaging analysis is invaluable for reducing the workload on human curators.

However, computational imaging analysis is not sufficiently accurate without human curatorial expertise. As software analyses pixels and bit data it may make errors in feature recognition which the curator must then correct. Analytical software is also limited to categorising what it is programmed to categorise, and if the code to recognise a particular feature is not present then the software will be unable to categorise certain images without the curator’s input. Human curators remain the most accurate at sorting images, though imaging software is invaluable at speeding up the process. The software is also incapable of carrying out contextual research for the images. This is the primary role fulfilled by the curator. Hence, curatorial expertise still has a central role in building functioning, informative image databases.

In conclusion, image curation today is the intersection between computational imaging analysis and traditional human curatorial methods. While computer software is beneficial and has a role to play in curation it is dependent on human curatorial expertise to provide contextual information and solve categorisation issues. Though computational imaging analysis is ubiquitous in everything from social media to smartphones, when it comes to image curation it is human curators who are the most accurate and necessary, even when the work of curation is undeniably tedious.

Leave a Reply