Crowdsourced cropping from the V&A: http://collections.vam.ac.uk/crowdsourcing
Generating Lots of Data Seems overwhelming! V&A has 120K images here! 30 sec/image = 120 images in 1 hr 1000 person hours This would require 5,000 people to each crop just 24 image or spend approx 12 min This is very doable
Steve in Action Funded in 2008 by the IMLS Led by the New Media Consortium in collaboration with IMA, Susan Chun and a host of museum partners A Few Project Goals Make Social Tagging Easy Develop Innovative NewInterfaces Facilitate Cross-CollectionSearch / Browsing
Steve in Action Features Simple Import (CSV, CDWA, Scraping) Hosted and Themable Data Collection Platform Powerful API Access Cut-n-Paste Tagging Widgets for Easy Integration
Tagcow Use crowdsourcing to add tags / data to image collections Cost $0.15 - $0.20 per image Tagcow uses software built on Amazon’s Mechanical Turk to process 100,000’s of images per day.
Mechanical Turk Demographics Source:PanosIpeirotis - http://behind-the-enemy-lines.blogspot.com/2008/03/mechanical-turk-demographics.html
IMA and Tagcow IMA gave Tagcow links to about 26,000 collection objects with images Tagcow returned 298,668 Total Tags 254,130 descriptive tags (28,708 distinct) 44,538 color tags Term Frequency: Min (1), Max(4299), Avg(8.85) Document Frequency: Min (1) Max(134) Avg(9.94) 29,174 tags with more than one word
So, 300,000 tags… can’t we just make a Wordleoutta that?
Funded in 2008 by IMLS Led by University of Maryland in collaboration with IMA, Susan Chun and a working group of museums. Studying the relationships between social tags, scholarly text and resources, and the application of trust networks to improve access to museum collections.