AViD Dataset: Anonymized Videos from Diverse Countries

NeurIPS 2020  ·  AJ Piergiovanni, Michael S. Ryoo ·

We introduce a new public video dataset for action recognition: Anonymized Videos from Diverse countries (AViD). Unlike existing public video datasets, AViD is a collection of action videos from many different countries. The motivation is to create a public dataset that would benefit training and pretraining of action recognition models for everybody, rather than making it useful for limited countries. Further, all the face identities in the AViD videos are properly anonymized to protect their privacy. It also is a static dataset where each video is licensed with the creative commons license. We confirm that most of the existing video datasets are statistically biased to only capture action videos from a limited number of countries. We experimentally illustrate that models trained with such biased datasets do not transfer perfectly to action videos from the other countries, and show that AViD addresses such problem. We also confirm that the new AViD dataset could serve as a good dataset for pretraining the models, performing comparably or better than prior datasets.

PDF Abstract NeurIPS 2020 PDF NeurIPS 2020 Abstract

Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Action Classification AViD SlowFast-101 16x8 Accuracy 50.9 # 2
Action Classification AViD I3D Accuracy 46.8 # 9
Action Classification AViD 2D ResNet-50 Accuracy 36.2 # 10
Action Classification AViD SlowFast-50 8x8 Accuracy 50.4 # 4
Action Classification AViD SlowFast-50 4x4 Accuracy 48.5 # 7
Action Classification AViD (2+1)D ResNet-50 Accuracy 48.8 # 6
Action Classification AViD RepFlow ResNet-50 Accuracy 50.5 # 3
Action Classification AViD Two-Stream 3D ResNet-50 Accuracy 50.1 # 5
Action Classification AViD 3D ResNet-50 Accuracy 48.2 # 8
Action Detection Charades 3D ResNet-50 + super-events pretrained on AViD mAP 25.2 # 8
Action Detection Charades 3D ResNet-50 pretrained on AViD mAP 23.2 # 12


No methods listed for this paper. Add relevant methods here