A Short Note on the Kinetics-700-2020 Human Action Dataset

Abstract

We describe the 2020 edition of the DeepMind Kinetics human action dataset, which replenishes and extends the Kinetics-700 dataset. In this new version, there are at least 700 video clips from different YouTube videos for each of the 700 classes. This paper details the changes introduced for this new release of the dataset and includes a comprehensive set of statistics as well as baseline results using the I3D network.

References in corpus (5)

• The Kinetics Human Action Video Dataset
• A Short Note about Kinetics-600
• The AVA-Kinetics Localized Human Actions Video Dataset
• Revisiting the Effectiveness of Off-the-shelf Temporal Modeling Approaches for Large-scale Video Classification
• Exploiting Spatial-Temporal Modelling and Multi-Modal Fusion for Human Action Recognition

Cited by in corpus (73)

• Learning Transferable Visual Models From Natural Language Supervision
• Transformers in Vision: A Survey
• Human Action Recognition from Various Data Modalities: A Review
• Perceiver IO: A General Architecture for Structured Inputs & Outputs
• Self-Supervised MultiModal Versatile Networks
• Video Transformers: A Survey
• BVI-DVC: A Training Database for Deep Video Compression
• MDMMT: Multidomain Multimodal Transformer for Video Retrieval
• SlowFast Networks for Video Recognition
• A Comprehensive Study of Deep Video Action Recognition
• The AVA-Kinetics Localized Human Actions Video Dataset
• Would Mega-scale Datasets Further Enhance Spatiotemporal 3D CNNs?
• Analyzing Human-Human Interactions: A Survey
• Zero-Shot Action Recognition in Videos: A Survey
• Less is More: ClipBERT for Video-and-Language Learning via Sparse Sampling
• Active Contrastive Learning of Audio-Visual Video Representations
• Measurement in AI Policy: Opportunities and Challenges
• Parameter Efficient Multimodal Transformers for Video Representation Learning
• VIMPAC: Video Pre-Training via Masked Token Prediction and Contrastive Learning
• Video Action Understanding
• Learn to cycle: Time-consistent feature discovery for action recognition
• UNIK: A Unified Framework for Real-world Skeleton-based Action Recognition
• Self-similarity Prior Distillation for Unsupervised Remote Physiological Measurement
• A Large-Scale Study on Unsupervised Spatiotemporal Representation Learning
• Omni-sourced Webly-supervised Learning for Video Recognition
• Action Genome: Actions as Composition of Spatio-temporal Scene Graphs
• MoViNets: Mobile Video Networks for Efficient Video Recognition
• Learning Video Representations from Textual Web Supervision
• Proposal Relation Network for Temporal Action Detection
• Deep Audio-Visual Learning: A Survey
• Transformation-based Adversarial Video Prediction on Large-Scale Data
• Three-Stream 3D/1D CNN for Fine-Grained Action Classification and Segmentation in Table Tennis
• Visual Relationship Detection using Scene Graphs: A Survey
• Actor-Context-Actor Relation Network for Spatio-Temporal Action Localization
• MER-GCN: Micro Expression Recognition Based on Relation Modeling with Graph Convolutional Network
• IMUTube: Automatic Extraction of Virtual on-body Accelerometry from Video for Human Activity Recognition
• DFEW: A Large-Scale Database for Recognizing Dynamic Facial Expressions in the Wild
• Contextual Emotion Recognition using Large Vision Language Models
• Learning from Temporal Gradient for Semi-supervised Action Recognition
• PERF-Net: Pose Empowered RGB-Flow Net
• Towards Training Stronger Video Vision Transformers for EPIC-KITCHENS-100 Action Recognition
• Uni-Perceiver: Pre-training Unified Architecture for Generic Perception for Zero-shot and Few-shot Tasks
• Object classification on video data of meteors and meteor-like phenomena: algorithm and data
• MAiVAR: Multimodal Audio-Image and Video Action Recognizer
• Scalable and Accurate Self-supervised Multimodal Representation Learning without Aligned Video and Text Data
• TinyAction Challenge: Recognizing Real-world Low-resolution Activities in Videos
• A Benchmark for Modeling Violation-of-Expectation in Physical Reasoning Across Event Categories
• Action Class Relation Detection and Classification Across Multiple Video Datasets
• 1st place solution for AVA-Kinetics Crossover in AcitivityNet Challenge 2020
• Global Semantic Descriptors for Zero-Shot Action Recognition
• Learning from Videos with Deep Convolutional LSTM Networks
• VidTr: Video Transformer Without Convolutions
• Spatial-Temporal Alignment Network for Action Recognition and Detection
• ASL Recognition with Metric-Learning based Lightweight Network
• Creating a Large-scale Synthetic Dataset for Human Activity Recognition
• Weakly-Supervised Temporal Action Localization Through Local-Global Background Modeling
• Alleviating Over-segmentation Errors by Detecting Action Boundaries
• Exploring Stronger Feature for Temporal Action Localization
• An Evaluation of Large Pre-Trained Models for Gesture Recognition using Synthetic Videos
• Efficient Action Recognition Using Confidence Distillation
• A Stronger Baseline for Ego-Centric Action Detection
• Contrastive Learning of Global-Local Video Representations
• From Recognition to Prediction: Analysis of Human Action and Trajectory Prediction in Video
• Machine-Generated Hierarchical Structure of Human Activities to Reveal How Machines Think
• Time and Frequency Network for Human Action Detection in Videos
• Multimodal Generation of Novel Action Appearances for Synthetic-to-Real Recognition of Activities of Daily Living
• Comparison of Spatiotemporal Networks for Learning Video Related Tasks
• Retrieving and Highlighting Action with Spatiotemporal Reference
• Efficient Modelling Across Time of Human Actions and Interactions
• Multi-Object Tracking with Hallucinated and Unlabeled Videos
• SegCodeNet: Color-Coded Segmentation Masks for Activity Detection from Wearable Cameras
• Beyond Short Clips: End-to-End Video-Level Learning with Collaborative Memories
• Unsupervised Few-Shot Action Recognition via Action-Appearance Aligned Meta-Adaptation