Vision-based human activity recognition: a survey
- PDF / 3,034,826 Bytes
- 47 Pages / 439.642 x 666.49 pts Page_size
- 34 Downloads / 232 Views
Vision-based human activity recognition: a survey Djamila Romaissa Beddiar1,3 Abdenour Hadid3
· Brahim Nini1 · Mohammad Sabokrou2 ·
Received: 10 June 2019 / Revised: 12 March 2020 / Accepted: 1 May 2020 / © The Author(s) 2020
Abstract Human activity recognition (HAR) systems attempt to automatically identify and analyze human activities using acquired information from various types of sensors. Although several extensive review papers have already been published in the general HAR topics, the growing technologies in the field as well as the multi-disciplinary nature of HAR prompt the need for constant updates in the field. In this respect, this paper attempts to review and summarize the progress of HAR systems from the computer vision perspective. Indeed, most computer vision applications such as human computer interaction, virtual reality, security, video surveillance and home monitoring are highly correlated to HAR tasks. This establishes new trend and milestone in the development cycle of HAR systems. Therefore, the current survey aims to provide the reader with an up to date analysis of vision-based HAR related literature and recent progress in the field. At the same time, it will highlight the main challenges and future directions. Keywords Human activity recognition · Behavior understanding · Action representation · Action detection · Computer Vision · Survey
1 Introduction Human activity recognition is often associated to the process of determining and naming activities using sensory observations [213]. More specifically, a human activity (HA) refers to the movement (s) of one or several parts of the person’s body. This can be either atomic or composed of many primitive actions performed in some sequential order. Therefore, human activity recognition should allow labeling the same activity with the same label even when Djamila Romaissa Beddiar
ad [email protected] 1
Research Laboratory on Computer Science’s Complex Systems, Larbi Ben M’hidi University, Oum El Bouaghi, Algeria
2
School of Computer Science, Institute for Research in Fundamental Sciences (IPM), Tehran Province, Iran
3
Center for Machine Vision Research, Computer Science and Engineering, University of Oulu, Oulu, Finland
Multimedia Tools and Applications
performed by different persons under different conditions or styles. HAR systems attempt to automatically analyze and recognize such HAs using the acquired information from the various types of sensors [4, 7]. Besides, HAR outputs can be employed to guide subsequent decision-support systems. For instance, authors in [103] proposed an HAR system that can help a teacher to control a multi-screen and multi-touch teaching tool, such as sweeping right or left to access the previous or next slide, call the eraser tool to rub out the wrong content, among others. Similarly, the work of [223] aims at ensuring a good implementation of various Human Computer interaction systems. To this end, the underlined HAR systems are generally preceded by an activity detection task. This consists of the tem
Data Loading...