Hogarth Challenge!
Tags: Art; Tech; Machine Learning; Hogarth; Gillray;
An interest in copies is a key theme that ties our various areas of research together. From the perspective of our ‘Art’ focus, copying of key images such as Hogarth’s Gin Lane or Gillray’s The Plumb-pudding in danger has been effective means to convey ideas, themes, and concepts. This reproduction of imagery in mediums such as cartooning gives rise to interesting questions:
- What makes an image so powerful? And why are William Hogarth’s works so effective?
- What forms of copying and borrowing have always existed in and around cartooning, and over decades of shifting copyright laws? What is “Hogarthomania” today?
- Can we chart the circulation of an image and its influence over time and through iterations?
From a tech perspective, the referencing of images by artists – often reproduced in their own unique visual style – poses interesting challenges. While computer vision is effective at identifying exact replicas of images, the field has given less attention towards identifying instances of referencing or ‘copying’ imagery.
In this way, frequently referenced works such as Gin Lane or The Plumb-pudding in danger provide useful case studies to train AI to recognise instances of referencing or copying. Similarly, an AI that is able to recognise the references can be useful for addressing questions such as the ones posed above. Our Twitter competition was thus directed towards this exciting nexus.
Announced at the second session of the 2021 ArTechLaw MineCraft Summit, the Twitter competition required participants to submit their own ‘copy’ of either Gin Lane or The Plumb-pudding in danger mentioning both @WilliamHogarth5 and @MartinRowson in their entry.
The Twitter account @WilliamHogarth5 has been linked with a Twitter bot running the ArtMiner algorithm which assessed the entries for spatially similar features, providing a percentage score of the image’s similarity.
Example: https://twitter.com/WilliamHogarth5/status/1417578403071504385?s=20&t=Wto6BySX43rab974HDE9rA
Over the course of a few months, we received a number of amazing entries - all of which can be viewed in this YouTube video and in June @MartinRowson announced the winners! One set of winners were announced based on the ArtMiner algorithm similarity score and another set of winners were decided by a panel of judges (Ronald Paulson, Tim Clayton, @blueloutter, Cynthia Roman & Cristina Sofia Martinez).
| Algorithm Winners | Panel of Judges Winners | ||||
|---|---|---|---|---|---|
| Gin Lane | |||||
| 1 | @nickymonkey | https://twitter.com/nickymonkey/status/1385827208439156736 | 1 | @Ybretsae | https://twitter.com/Ybretsae/status/1397982062422994948/photo/1 |
| 2 | @bainescartoons | https://twitter.com/bainescartoons/status/1395469573017714688/photo/1 | 2 | @Trumpton | https://twitter.com/Trump_ton/status/1386965130651525120/photo/1 |
| 3 | @GlasgowRiot | https://twitter.com/GlasgowRiot/status/1400813284773797888/photo/1 | 3 | @t_colgate | https://twitter.com/t_colgate/status/1385630096497725443 |
| Plum-pudding | |||||
| 1 | @cartoonsidrew | https://twitter.com/cartoonsidrew/status/1386403987201478657/photo/1 | 1 | @Coldwar_Steve | https://twitter.com/Coldwar_Steve/status/1390290926304591878/photo/1 |
| 2 | @t_colgate | https://twitter.com/t_colgate/status/1386033772928421889 | 2 | @nickymonkey | https://twitter.com/nickymonkey/status/1397456317871271936/photo/1 |
| 3 | @DavidWaywell | https://twitter.com/DavidWaywell/status/1394239223478239233/photo/1 | 3 | @JamesDFMellor | https://twitter.com/JamesDFMellor/status/1386641687120908289/photo/1 |
While the competition is now closed, you can still Tweet your Gin Lane and Plum-Pudding copies to @WilliamHogart5 to get a similarity score (and help our research). If you do submit a copy, make sure you let the Twitter bot know what image to compare by using #DrawGinLane/#DrawThePlumPudding