Social media images can help create 3D reconstructions of cultural heritage. Social media, specifically Twitter, are digital repositories of images key to preservation efforts of monuments. Much like museums, visitors use Twitter and share photos of artefacts and monuments they encounter actively promoting them. In this instance, promotion can serve as preservation. Today we discuss a paper that explains how to identify content, filter images and provide 3D models using Structure from Motion algorithm from Twitter.
Digital access and virtual heritage are growing in the cultural sector as they provide meaningful communication, promotion and conservation. Large museums more often employ digitisation and virtual galleries for visitors to experience. It is safe to assume that these institutions do not crowdsource data for their collections. Small museums, unable to afford digitisation, can begin digitally preserving their assets with Twitter’s help. Therefore, minimising the gap between digital services and conservation between cultural institutions.
How it works
The study revolves around Twitter images, as it’s short text format allows for easier content identification. Since not all photos on Twitter relate to cultural heritage, the study establishes criteria and selection tools. The concept develops in 3 areas :
1)Identify and form characteristics
Based on the text content of tweets, an event detection algorithm runs along geo-tags and timeframe. Using the statistical measure TF–IDF score evaluating how important a word is in a text document according to its frequency. Additionally, the number of tweets, retweets and followers are taken into account.
2)Collect images
The algorithm collects images according to the characteristics set above and filters them into groups removing the outliers.
3)Create 3D-models
The Structure from Motion algorithm calculates 3D structures from 2D images each image serves as a point to reconstruct the model with the analysis of corresponding images and outlier removal.
The drawback of this image gathering format is that sometimes details and the complete structure of a monument are not evident, for instance, it would be difficult to recreate an entire archaeological site solely based on social media. Besides, this method accelerates the time required for rendering 3D models and is cost-efficient.
The study goes in-depth on the parameters and tools for data collection, the reasoning behind the decisions. For this article, we summarised and simplified their process, to give a general understanding of the concepts at hand.
Our take
This case study offers innovative applications in various heritage management areas, particularly beneficial to small museums and communities. Communication is now an open two-way channel, “traditionally” museums share images, then retweeted by the audience here the audience participates in the documentation of objects and monuments. Moreover, this method reduces the costs of digitisation (equipment, photographers etc.) and time. Additionally, digital conservation becomes readily available to all heritage, covering the increasing demand due to climate change, COVID-19 pandemic and lack of funding. Furthermore, this practice is effectively Smart Heritage and can help integrate heritage in Smart City policy as utilises both new technologies and public participation. The above are some of the multiple applications of rendering 3D models through social media images, for Local Approach the most valuable element is its participatory and open character, everyone’s visit matters.
I want to learn more:
Automatic 3D Modeling and Reconstruction of Cultural Heritage Sites from Twitter Images by Anastasios Doulamis, Athanasios Voulodimos, Eftychios Protopapadakis,Nikolaos Doulamis and Konstantinos Makantasis
or
Automatic 3D Modeling and Reconstruction of Cultural Heritage Sites from Twitter Images by Anastasios Doulamis, Athanasios Voulodimos, Eftychios Protopapadakis,Nikolaos Doulamis and Konstantinos Makantasis
