The AVA-Kinetics Localized Human Actions Video Dataset

TL;DR

The AVA-Kinetics dataset extends the original AVA dataset with over 230,000 clips from Kinetics-700, annotated for human actions, enabling improved action classification research with baseline evaluations.

Contribution

This paper introduces the AVA-Kinetics dataset, combining Kinetics-700 videos with AVA annotations, and provides baseline results using the Video Action Transformer Network.

Findings

Baseline evaluation shows improved action classification performance.

The dataset contains over 230k annotated clips.

Provides comprehensive statistics and annotation process details.

Abstract

This paper describes the AVA-Kinetics localized human actions video dataset. The dataset is collected by annotating videos from the Kinetics-700 dataset using the AVA annotation protocol, and extending the original AVA dataset with these new AVA annotated Kinetics clips. The dataset contains over 230k clips annotated with the 80 AVA action classes for each of the humans in key-frames. We describe the annotation process and provide statistics about the new dataset. We also include a baseline evaluation using the Video Action Transformer Network on the AVA-Kinetics dataset, demonstrating improved performance for action classification on the AVA test set. The dataset can be downloaded from https://research.google.com/ava/