Our monthly webinar archive

Giles presents some of the ways that cultural heritage collections are using computer vision (or visual AI) for collections management and research, focussing particularly on the work of the Oxford Visual Geometry Group and its collaborators. The talk will demo such applications as visual search, image classification, object detection, deduplication and image registration.

Presenter bio

Giles Bergel is Senior Researcher in Digital Humanities in the Department of Engineering Science at the University of Oxford, with additional appointment as Digital Project Officer at the History of Science Museum in Oxford. A book historian by training, he broadly works on digitisation projects for museums and libraries, particularly those involving computer vision.

In this online lecture, Dale Moulding, light microscope specialist working at UCL in London, will be talking about his experience working on Image analysis applied to the fields of archaeometry and cultural heritage. The talk will start with an introduction to the Fiji / ImageJ software, a description of what digital images are and insights into basic image processing steps.

Using two study cases as examples:

• The application of image analysis to automate fossilised Millet seed measurements. 

• Fiji as a tool to measure potsherds morphometrics for post-depositional erosion and fragmentation,

Dale will show how image analysis using a tool like Fiji helped him unravel some mysteries about the history of the archaeological artefacts investigated and the life of the people who used them.

Following this presentation (if time), we will have Q&A and breakout rooms for networking!

Presenter bio

My name is Dale Moulding and I am a light microscope specialist at the GOS Institute of Child Health in London (UK). My background (16 years research) is in immunology and cell biology, with a heavy emphasis to applying microscopy to understand immune processes. In the last 5 years, I have been running the confocal and light microscopy core facility at my institute where I train users in the most appropriate imaging system for their research and provide custom image analysis protocols. As part of this, I run an open access image analysis course through which I have helped analyse mostly biological samples, but also ice-cream, seaweed, paintings, persian rugs & archaeological samples.

To know get more information about the facility and the course, follow the links below:
https://www.ucl.ac.uk/ich/core-scientific-facilities-centres/confocal-microscopy(facility)
https://www.ucl.ac.uk/child-health/research/core-scientific-facilities-centres/confocal-microscopy/training-courses). (course)

Further readings

-ImageJ documentation: https://imagej.nih.gov/ij/

-Publications :

Vindrola-Padrós, B, et al. (2019). Working with broken agents: Exploring computational 2D morphometrics for studying the (post)depositional history of potsherds. Journal of Archaeological Science, 104, 19–33.

Saunders, D., Collmann, R., & Borda, A. (2017). Reflectance Transformation Imaging and ImageJ: Comparing Imaging Methodologies for Cultural Heritage Artefacts, 350–357.

Cámara, B., et al. (2017). Biodeterioration of marble in an underwater environment. Science of the Total Environment, 609(September), 109–122.

For this activity, we will be using a temperature and humidity dataset and how to use R to download, perform some summaries and graph straight into a report. If you would like to follow along, please do go ahead and download R & R Studio (free version) and the dataset we will be working with is here. I had a play around with this dataset last year and the output pdf is here.

See here for further information about learning to code in R: https://www.concode.info/learning-code

Digital archives reflect the physical embodiment of the institution and/or its organisational culture over and above the needs of the user.DCD Labs, a collaborative project with Dance Collection Danse is turning this around. It puts the user needs at the heart of all it does. DCD Discover (https://discover.dcd.ca/items) is the result, built by the user for the user.Gavin, Director of Dgen, will describe his vision for ‘digital archives’, and the importance of digital-first and for building for people not things. ‘We never start with code. We ask what people want first.' Gavin Starks, July 2020.We will take a tour of DCD Discover, giving you the opportunity to add your archives, comment and join in this user designed iterative process.

Looking for a new way to connect with your stakeholders? Zooniverse is a free platform that enables citizen science projects across a range of disciplines, including cultural heritage. In this presentation, Emily will show us how to create a project in Zooniverse, and clean the data in R to prepare for analysis. The presentation will be based on Heritage Combs, a group project created during the Crowd-Sourced and Citizen Data for Cultural Heritage module at UCL.

Have you ever spent far too long trying to clean up messy data in Excel? In this presentation, Jen will go over some of the amazing applications of Google's free online application, OpenRefine. OpenRefine is particularly well suited to cleaning up text-based data, but also has some date-based and numerical data cleaning capabilities. This workshop is relevant to everyone at any stage of their data science learning.

Thanasis is part of the Linked Conservation Data project (LCD), funded by the AHRC. LCD is a consortium of partners working on sharing conservation records using Linked Open Data. The project focuses on guidelines for sharing conservation vocabularies and establishing a common model for expressing conservation data.

Justin is part of the Collection Management unit at the Smithsonian’s National Air and Space Museum, specializing in conservation and with a lot of experience moving collections. Justin has completed certifications in T-SQL and SQL server and is currently trying to learn VBA and tighten up his Access database skills.

Nathan takes us through the basics about Git and working with Github

Josep takes us trough the basics on data science in the heritage field. While Savannah and Vincent talk about their translation from an Excel tool for Temperature and Relative Humidity analysis to code in R.