Program

Time Talk Slides Memories
14:00 Bernard Ghanem: Opening remarks Slides
14:15 Joao Carreira: Video Understanding and the Kinetics Dataset Slides
14:45 Jitendra Malik: AVA: A Video Dataset of Atomic Visual Actions Slides
15:15 Huijuan Xu: R-C3D: Region-Convolutional 3D Network for Activity Detection (Most Innovative) Slides
15:35 Bernard Ghanem: Task1: Untrimmed Video Classification (Dataset and Results) Slides
15:45 Afternoon Break
16:15 Jiankang Deng: Improve Untrimmed Video Classification and Action Localization by Human and Object Tubelets (Winner Task 1) Slides
16:35 Brian Zhang: Task2: Trimmed Action Recognition (Dataset and Results) Slides
16:45 Chuang Gan: Revisiting the Effectiveness of Off-the-shelf Temporal Modeling Approaches for Video Recognition (Winner Task 2) Slides
17:05 Fabian Caba: Task3&4: Temporal Action Proposals and Localization (Dataset and Results) Slides
17:20 Zheng Shou: Temporal Convolution Based Action Proposal: Submission to ActivityNet 2017 (Winner Task3&4) Slides
17:40 Ranjay Krishna: Task5: Dense-Captioning Events in Videos (Dataset and Results) Slides
17:50 Ting Yao: ActivityNet Challenge 2017: Dense-Captioning Event in Videos (Winner Task 5) Slides
16:10 Bernard Ghanem: Closing Remarks Slides

Important Dates

May 1, 2017: Development kit and data will be available.
June 12, 2017: Test subsets will be available.
June 19, 2017: Evaluation server will be available.
July 5, 2017: Leaderboard will be public.
July 9, 2017 July 16, 2017 (11:59 PM Hawaii-Aleutian Time Zone): Challenge submission deadline.
July 11, 2017 July 18, 2017 (11:59 PM Hawaii-Aleutian Time Zone): Paper submission deadline.
July 12, 2017 July 19, 2017 (11:59 PM Hawaii-Aleutian Time Zone): Winners announcement.
July 26, 2017: Workshop session.