In the age of Web 2.0 and social media and online social networking, user-generated content has become the main source of information for many people. Nowhere is this truer than on popular photo-sharing websites. While the likes of Instagram have moved to the fore, there are still millions of people using the much older and more sophisticated service Flickr. This site was established in 2004 by Ludicorp, it was then bought by Yahoo, which itself was acquired by Verizon, and the Flickr component sold on to another photo site, SmugMug in April 2018.
Research published in the International Journal of Information Technology and Management describes an empirical study on how tags might be mined from the comments Flickr users make on each other’s photographs, videos, and other images. Haijun Zhang, Jingxuan Li, and Bin Luo of the Department of Computer Science at Harbin Institute of Technology Shenzhen Graduate School and Yan Li of the School of Computer Engineering at Shenzhen Polytechnic in Shenzhen, both in Guangdong Province, China, hope to develop a technique that might be useful in the curation of the huge database of images stored by Flickr.
They have developed a two-phase approach wherein tags are generated from comments on a given photo and then these are ranked.
“In the phase of candidate tags generation, two methods are introduced relying on natural language processing (NLP) techniques, namely word-based and phrase-based,” the team explains. “In ranking and recommending tags, we proposed an algorithm by jointly modelling the location information of candidate tags, statistical information of candidate tags and semantic similarity between candidate tags. Extensive experimental results demonstrate the effectiveness of our method.”
The team claims this to be the first paper addressing the problem of tagging photos in the Flickr database in this way, it could assist in curating the collection, especially where photos have not been tagged initially by the person that uploads them.
Zhang, H., Li, J., Luo, B. and Li, Y. (2019) ‘Needle in a haystack: an empirical study on mining tagsfrom Flickr user comments’, Int. J. Information Technology and Management, Vol. 18, Nos. 2/3, pp.297–326.