Results of the 1st Asynchronous CASTLE Challenge at the Joint Egocentric Vision Workshop in Conjunction with CVPR 2026
Authors: Luca Rossetto, Werner Bailer, Cathal Gurrin, Graham Healy, Omar Shahbaz Khan, Stevan Rudinac, Klaus Schöffmann, Allie Tran
Organizations: Dublin City University Dublin, Ireland · JOANNEUM RESEARCH Graz, Austria · IT University of Copenhagen Copenhagen, Denmark · University of Amsterdam Amsterdam, The Netherlands · Klagenfurt University Klagenfurt, Austria
This report summarizes the contributions and results of the 1st Asynchronous CASTLE Challenge at the Joint Egocentric Vision Workshop in conjunction with CVPR 2026.
Figures & tables
Team
Correct Answers
Accuracy
WDL
108
0.584
MARS
105
0.568
TAHAKOM
101
0.546
CuriosAI (SVA)
92
0.497
CuriosAI (TMKG)
65
0.351
Table 1: Official Challenge Results
Teams answered correctly
Number of Questions
0
9
1
34
2
48
3
47
4
35
5
12
Table 2: Aggregate histogram over correctly answered questions and teams.
Egocentric video understanding studies how AI interprets first-person video to recognize actions, track interactions, reason about space and time, and assist users.