First Leaderboard
For the first initial validation set, each image had multiple renderings, namely the ones submitted by participants. Subsequently, all possible pairs of these renderings were generated and presented to 20 evaluators via the Yander Tasks platform. Their task was to indicate their preference between the images in each pair. This process was iterated for every image in the validation set. The submission's score was determined by the performance of its images in these pairwise evaluations.Rank | Team | Member(s) | Affiliation | Score mean |
---|---|---|---|---|
1 | MiAlgo (submission II) | Shuai Liu | Xiaomi Inc., China | 0.8003 |
- | baseline | - | Organizers Team | 0.5810 |
2 | OzUVGL (submission I) | Furkan Kınlı, Doğa Yılmaz, Barış Özcan, Furkan Kıraç | Özyeğin University Bahçeşehir University |
0.5017 |
3 | OzUVGL (submission II) | Furkan Kınlı, Doğa Yılmaz, Barış Özcan, Furkan Kıraç | Özyeğin University Bahçeşehir University |
0.2391 |
4 | MiAlgo (submission I) | Shuai Liu | Xiaomi Inc., China | 0.2349 |