A Short Note about Kinetics-600

Abstract

We describe an extension of the DeepMind Kinetics human action dataset from 400 classes, each with at least 400 video clips, to 600 classes, each with at least 600 video clips. In order to scale up the dataset we changed the data collection process so it uses multiple queries per class, with some of them in a language other than english -- portuguese. This paper details the changes between the two versions of the dataset and includes a comprehensive set of statistics of the new version as well as baseline results using the I3D neural network architecture. The paper is a companion to the release of the ground truth labels for the public test set.

arXiv commentCompanion to public release of kinetics-600 test set labels

References in corpus (4)

• UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
• The Kinetics Human Action 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 (77)

• Is Space-Time Attention All You Need for Video Understanding?
• Human Action Recognition from Various Data Modalities: A Review
• A Short Note on the Kinetics-700-2020 Human Action Dataset
• VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text
• A Review on Deep Learning Techniques for Video Prediction
• Self-Supervised MultiModal Versatile Networks
• Drop an Octave: Reducing Spatial Redundancy in Convolutional Neural Networks with Octave Convolution
• Adversarial Video Generation on Complex Datasets
• SlowFast Networks for Video Recognition
• Thinking Fast and Slow: Efficient Text-to-Visual Retrieval with Transformers
• A Comprehensive Study of Deep Video Action Recognition
• Spatiotemporal Contrastive Video Representation Learning
• Predicting post-operative right ventricular failure using video-based deep learning
• Multiscale Vision Transformers
• From CNNs to Transformers in Multimodal Human Action Recognition: A Survey
• Medical Imaging and Machine Learning
• Zero-Shot Action Recognition in Videos: A Survey
• Measurement in AI Policy: Opportunities and Challenges
• Sign Language Recognition via Skeleton-Aware Multi-Model Ensemble
• Video Action Understanding
• CCVS: Context-aware Controllable Video Synthesis
• Action Machine: Rethinking Action Recognition in Trimmed Videos
• Skeleton Aware Multi-modal Sign Language Recognition
• Predicting Video with VQVAE
• RadFusion: Benchmarking Performance and Fairness for Multimodal Pulmonary Embolism Detection from CT and EHR
• Disentangling Label Distribution for Long-tailed Visual Recognition
• StNet: Local and Global Spatial-Temporal Modeling for Action Recognition
• Wisdom of Committees: An Overlooked Approach To Faster and More Accurate Models
• AViD Dataset: Anonymized Videos from Diverse Countries
• A Large-Scale Study on Unsupervised Spatiotemporal Representation Learning
• Metric-Based Few-Shot Learning for Video Action Recognition
• MoViNets: Mobile Video Networks for Efficient Video Recognition
• Deep Audio-Visual Learning: A Survey
• Learning Video Representations from Textual Web Supervision
• Temporal Attentive Alignment for Large-Scale Video Domain Adaptation
• Revisiting 3D ResNets for Video Recognition
• Continual 3D Convolutional Neural Networks for Real-time Processing of Videos
• Scaling Autoregressive Video Models
• Transformation-based Adversarial Video Prediction on Large-Scale Data
• Temporal Interlacing Network
• Weakly Supervised Attention Model for RV StrainClassification from volumetric CTPA Scans
• Learning to Compress Videos without Computing Motion
• HACS: Human Action Clips and Segments Dataset for Recognition and Temporal Localization
• Learning from Label Relationships in Human Affect
• Reconfigurable Cyber-Physical System for Lifestyle Video-Monitoring via Deep Learning
• AttentionNAS: Spatiotemporal Attention Cell Search for Video Classification
• Shifted Chunk Transformer for Spatio-Temporal Representational Learning
• Self-Supervised Learning via multi-Transformation Classification for Action Recognition
• Optimization Planning for 3D ConvNets
• VGGSound: A Large-scale Audio-Visual Dataset
• Three Branches: Detecting Actions With Richer Features
• Zero-Shot Learning with Knowledge Enhanced Visual Semantic Embeddings
• A Study of Few-Shot Audio Classification
• Elaborative Rehearsal for Zero-shot Action Recognition
• Augmented 2D-TAN: A Two-stage Approach for Human-centric Spatio-Temporal Video Grounding
• Exploring Stronger Feature for Temporal Action Localization
• Temporal Fusion Network for Temporal Action Localization:Submission to ActivityNet Challenge 2020 (Task E)
• Multi-Label Activity Recognition using Activity-specific Features and Activity Correlations
• An Evaluation of Large Pre-Trained Models for Gesture Recognition using Synthetic Videos
• Efficient Action Recognition Using Confidence Distillation
• Top-1 Solution of Multi-Moments in Time Challenge 2019
• ParamCrop: Parametric Cubic Cropping for Video Contrastive Learning
• Hierarchical Video Generation for Complex Data
• Recent Progress in Appearance-based Action Recognition
• Machine-Generated Hierarchical Structure of Human Activities to Reveal How Machines Think
• GCF-Net: Gated Clip Fusion Network for Video Action Recognition
• PGT: A Progressive Method for Training Models on Long Videos
• DeepLandscape: Adversarial Modeling of Landscape Video
• Comparison of Spatiotemporal Networks for Learning Video Related Tasks
• Three-stream network for enriched Action Recognition
• Learning Goals from Failure
• SegCodeNet: Color-Coded Segmentation Masks for Activity Detection from Wearable Cameras
• Controllable Attention for Structured Layered Video Decomposition
• Efficient Modelling Across Time of Human Actions and Interactions
• CTM: Collaborative Temporal Modeling for Action Recognition
• A Prospective Study on Sequence-Driven Temporal Sampling and Ego-Motion Compensation for Action Recognition in the EPIC-Kitchens Dataset
• Discriminative Video Representation Learning Using Support Vector Classifiers