A survey of video datasets for human action and activity recognition

By José M. Chaquet, Enrique J. Carmona, Antonio Fernández-Caballero. 2013

466 citations recorded by the source corpus.

Highlights? Description of datasets for video-based human activity and action recognition. ? 68 datasets reported: 28 for heterogeneous and 40 for specific human actions. ? Useful data, such as web for dowloading, published works or ground truth, are provided. ? Datasets are compared and classified. Vision-based human action and activity recognition has an increasing importance among the computer vision community with applications to visual surveillance, video retrieval and human-computer interaction. In recent years, more and more datasets dedicated to human action and activity recognition have been created. The use of these datasets allows us to compare different recognition systems with the same input data. The survey introduced in this paper tries to cover the lack of a complete description of the most important public datasets for video-based human activity and action recognition and to guide researchers in the election of the most suitable dataset for benchmarking their algorithms.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

RGB-D-based action recognition datasets: A survey (2016, background): “Review papers categorizing and summarizing their characteristics are available to help researchers in evaluating their algorithms [5, 6, 7].”
3D Human Activity Recognition with Reconfigurable Convolutional Neural Networks (2014, methodology): “Some compositional methods [33, 3] attempt to model activities by representing videos as sequences of ﬁxed length temporal segments.”
Evaluation of video activity localizations integrating quality and quantity measurements (2014, background): “The available datasets have been extensively studied in a recent survey [18].”
Category-Blind Human Action Recognition: A Practical Recognition System (2015, background): “Existing techniques for single-person action recognition are extensively surveyed in [2, 3, 8] with the majority of such methods using monocular RGB videos [9, 20, 21, 26].”
Human Action Recognition From Various Data Modalities: A Review (2020, background): “Specifically, since videos contain the temporal dynamics of human motions that are often crucial for HAR, most of the existing works focused on using videos to handle HAR [37], and only a few approaches utilized static images [38], [39], [40].”
Action Recognition Using Rate-Invariant Analysis of Skeletal Shape Trajectories (2016, methodology): “Rather than covering all ideas exhaustively, we direct interested readers to some recent surveys [5], [6], [7], [9] that together overview this research area.”
UniMiB SHAR: A Dataset for Human Activity Recognition Using Acceleration Data from Smartphones (2016, methodology): “[14] , are examples of datasets including samples from cameras, while the works by Wang et al.”
A Critical Review of Action Recognition Benchmarks (2013, background): “Two very recent examples include [3, 12] and [19].”
A Review on Human Activity Recognition Using Vision-Based Method (2017, background): “There have been a published good survey [4] which presents 22 Journal of Healthcare Engineering the available important public datasets; however, it mainly focused on the conventional RGB-based datasets and missed current depth-based datasets.”
Human action recognition using genetic algorithms and convolutional neural networks (2016, background): “proposed in the literature [8] to address different types of problems like recognition of realistic activities, interaction and multiview analysis from varying sources.”
Dynamic-MUSIC: accurate device-free indoor localization (2016, methodology): “Different technologies have been employed for indoor localization, such as camera [3, 4], RF [2, 33, 36, 37, 50], infrared [8, 13], Ultra-wideband (UWB) [9], sound [14, 41], and recently visible light [15, 20], etc.”
The Language of Actions: Recovering the Syntax and Semantics of Goal-Directed Human Activities (2014, methodology): “The number of stages was determined by cross validation using n = [3, 5, 10, 15] fixed stages.”
Online human gesture recognition from motion data streams (2013, background): “We set G = [5, 10, 15, 20] respectively when λ = 100 and β = −0.”
A Review of Human Activity Recognition Methods (2015, methodology): “A complete survey, which covers important aspects of human activity recognition datasets, was introduced by Chaquet et al. (2013). An appropriate human activity dataset is required for the development of a human activity recognition system.”
Fuzzy human motion analysis: A review (2014, background): “Please ref r to [37] for a complete list of the currently available datasets in HMA.”
Computer Vision and Image Understanding Computer Vision for Assistive Technologies (unknown): “Additional researches focus on the evidence that activities are normally related to the concept of interaction between a person with one or more people or between one or more people with objects of the surrounding environment ( Chaquet et al., 2013 ).”